<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Vytas Prazarkevicius on Medium]]></title>
        <description><![CDATA[Stories by Vytas Prazarkevicius on Medium]]></description>
        <link>https://medium.com/@vytotas?source=rss-172fc8740d10------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*8gQAigIrLbPC3kNG-eFjKw@2x.jpeg</url>
            <title>Stories by Vytas Prazarkevicius on Medium</title>
            <link>https://medium.com/@vytotas?source=rss-172fc8740d10------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 21:41:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@vytotas/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Grafana Dashboards for Windows Servers]]></title>
            <link>https://medium.com/geekculture/grafana-dashboards-for-windows-servers-80697d59d1dd?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/80697d59d1dd</guid>
            <category><![CDATA[grafana]]></category>
            <category><![CDATA[monitoring]]></category>
            <category><![CDATA[prometheus]]></category>
            <category><![CDATA[observability]]></category>
            <category><![CDATA[windows-server]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Sat, 17 Apr 2021 17:03:02 GMT</pubDate>
            <atom:updated>2021-04-17T17:03:02.603Z</atom:updated>
            <content:encoded><![CDATA[<h4>Understanding two key server performance metrics</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*YZ9KNVbQ49brmdCF" /><figcaption>Photo by <a href="https://unsplash.com/@ibrahimboran?utm_source=medium&amp;utm_medium=referral">Ibrahim Boran</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Prometheus and Grafana are becoming de-facto standards for monitoring linux servers. You will find countless guides of how to get Prometheus with Grafana running under 10 minutes. This includes beautiful, interactive dashboards out-of-the-box.</p><p>But what about monitoring Windows servers? The how-to guide space for setting up Windows with Prometheus and Grafana is much more empty. The good news is that you can still monitor your windows servers with Prometheus and Grafana. The bad news is that the variety of available metrics and Grafana dashboards out of the box is quite limited.</p><p>But don’t despair. In this how-to guide, I will introduce you to the most popular Grafana dashboard for Windows servers. You will get an overview of how to read the dashboard data about CPU and memory.</p><h3><strong>Dashboards</strong></h3><p>Typical post about Grafana focuses on setting up data sources, importing dashboards and tweaking “look and feel” of the panes. Instead of introducing you to the technical setup part, I will describe how to benefit from Grafana dashboards. At the end, understanding what a dashboard tells you about the performance is the most valuable part.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ddgBlyv7fblxxsyAIrZfgA.png" /><figcaption>Grafana dashboard</figcaption></figure><p>The most popular Grafana dashboard for Windows servers is called “Windows Node” and can be downloaded from Grafana website <a href="https://grafana.com/grafana/dashboards/2129">here</a>. This dashboard is compatible with WMI exporter 0.12 which you can download <a href="https://github.com/prometheus-community/windows_exporter/releases/tag/v0.12.0">here</a> and install by following <a href="https://github.com/prometheus-community/windows_exporter/tree/v0.12.0">instructions</a> on the Github page.</p><p>“Windows Node” Grafana dashboard contains many panels with various performance aspects of your windows server. CPU and memory are by far the most import server resources. In order to undestand those, we need to dig deeper into how CPU and memory is managed in Windows.</p><h4>CPU</h4><p>The CPU panel in the Grafana dashboard looks like that:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*56x0Ay54vjYPanThpHcXmw.png" /><figcaption>CPU Load panel</figcaption></figure><p>This panel shows what kind of work CPU has been doing for a period of time.</p><p>On the left side you see the Y axis with a top value of 4. In this particular example the monitored server has a single processor with 4 cores (e.g. 4 logical processors) and hence the max value of Y axis is 4.</p><p>On the right side of the panel we see the legend representing 5 different activities of the processor: <em>dcp, idle, interrupt, privileged, user</em>.</p><p>Let’s take a look at the legend closer:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/446/1*yl4UGvBYF7j9DDgYr28LWQ.png" /><figcaption>CPU Load panel legend</figcaption></figure><p>The other name for these activities is “processor modes”. When we say “processor is in the mode X” we mean that the processor is doing work of type X.</p><p>Let’s jump into definitions of those modes.</p><p><em>User mode</em> represents the work done for applications and different subsystems. The typical applications we run on servers are executed in <em>user mode</em>. If your application is computation-hungry, you will see user mode metric spiking up.</p><p><em>Privileged mode</em> is a work the processor does for a Windows operating system. When a Windows system service is called, the service will often run in <em>privileged mode</em> to gain the access to the system-private data. Such data is protected from the access of threads executing in <em>user mode</em>.</p><p><em>Interrupt mode</em> is a type of work done for Windows system to handle hardware requests. High level of interrupts may indicate a hardware or driver problem.</p><h4>Memory</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2j5WhTdamPQg-4VPFXA9Kg.png" /><figcaption>Memory panel legend</figcaption></figure><p>The memory pane in the dashboard displays information about three key metrics on server memory:</p><ul><li>virtual memory</li><li>physical memory</li><li>free physical memory</li></ul><p>Physical memory is the memory installed on a server (RAM). Processes pf applications do not access physical memory directly — only operating system does. Instead those processes use virtual memory which is created by the operating system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6a_CQFH15GETpBvEkzn6Tw.png" /></figure><p>Why do we need virtual memory? Processes sometimes require more memory than there is physical memory available. In such cases the operating system loads the processes sequentially to the physical memory. It waits for the end of each process and then it loads the next one. This switching repeats for all processes which are using virtual memory.</p><p>To do so, the operating system uses a technique called <em>Paging</em>. When the server does not have enough of physical memory for all processes, parts of information from physical memory is saved to a disk file called P<em>age File</em>.</p><h3>Summing up</h3><p>You should now have an understanding of how to read CPU and memory consumption charts. These metrics cover 80% of your needs when assessing performance bottlenecks of your servers (great illustration of Pareto principle).</p><p>If you need to dive-deep into observing how your application impacts CPU and memory, take a look at <a href="https://prometheus.io/docs/instrumenting/clientlibs/">Prometheus client libraries</a>. You can add those libraries directly to your application to collect application run-time metrics. Run-time metrics are usually helpful to answer “why” questions of your investigations of performance bottlenecks.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=80697d59d1dd" width="1" height="1" alt=""><hr><p><a href="https://medium.com/geekculture/grafana-dashboards-for-windows-servers-80697d59d1dd">Grafana Dashboards for Windows Servers</a> was originally published in <a href="https://medium.com/geekculture">Geek Culture</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Organizations Can Add New Tech and Avoid Developer Burnout]]></title>
            <link>https://medium.com/swlh/how-organizations-can-add-new-tech-and-avoid-developer-burnout-11ab68a3bad6?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/11ab68a3bad6</guid>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[developer-experience]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[platform]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Sat, 06 Mar 2021 15:44:43 GMT</pubDate>
            <atom:updated>2021-03-06T15:44:43.266Z</atom:updated>
            <content:encoded><![CDATA[<h4>Internal platforms as a means to maximize flow</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*PiyPZZrb6Lq7K1Je" /><figcaption>Photo by <a href="https://unsplash.com/@milkandbourbons?utm_source=medium&amp;utm_medium=referral">Dean Brierley</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Any new technology adoption impacts cognitive load of development teams. Cognitive load could be explained as the amount of mental energy required to perform a task. Too much of <em>cognitive load </em>can create overload for developers. It then causes slower <em>flow of change</em> and delivery bottlenecks. It can also increase likelihood of human errors, add frustration and stress or worse -induce burnout.</p><p>How do new technologies come to lives of developers an how does it impact their cognitive load? And what can we do <em>practically</em> to better manage unwanted effects of new technology to developer teams?</p><p>In this post I am going to look at the example of Kubernetes adoption in a typical team. I will discuss the mental overload teams face and how such overload could be addressed in organizations.</p><h4>Costs of adding new technology</h4><p>New technologies often benefit one or more IT teams and puts more complexity for the other team without visible benefit for the latter. Let’s take Kubernetes as an example. Kubernetes is a container orchestration framework. It puts container deployment, operating and scaling to another level for Operations teams. Without such a container orchestrator Operations teams would need to use custom automation, deployment, networking and cluster management scripts.</p><p>Kubernetes is a new(-ish) technology which targets many types of users. Cluster admins, infrastructure teams and advanced users use command line tools and scripts to manage infrastructure and applications using it.</p><p>Kubernetes is not developer-friendly. As it aims to be everything for everyone, it has lot of complexity which is not encapsulated well enough.</p><p>So what do developers get from using Kubernetes? Sure, deployments are more reliable and more automated. But this brings more benefits to Ops teams, right? And at what cost? Developers need to wrap their minds around pods, replicas, nodes, health checks, helm charts, command-line deployment tools and more.</p><p>Why should developers care about deployments at first place? Why should they care how applications behave in production? But these are “simple” questions which contradict “DevOps culture” and should not be asked openly.</p><h4>Focus on value delivered</h4><p>Now let’s explore the idea of “developers create value to users by adding/changing application code”. Not every code is valuable. Not every feature brings value to users. There is a big topic of outputs vs outcomes (e.g. features delivered vs. user value added) which I won’t touch here.</p><p>Relevant part from that big topic is a notion of feedback loops (especially user feedback). We cannot really always know if a feature or even a finished product will be valuable upfront. Marty Cagan calls it as one of two inconvenient truths of product management. According to him, at least half of our ideas about new products or features aren’t going to work. In Scrum methodology, this is popularized by the 80:20 rule.</p><p>Therefore developers need to receive feedback which confirms that the value they created has reached the users. Value delivery can fail at many levels. From an application crashing to users not knowing how to use a released feature. Thus if developers care that their work brings value to a user, they need constant value delivery feedback.</p><p>Such feedback may take many forms. Feedback that an application got deployed correctly, that it is running as expected, that users use that feature as intended, etc.</p><p>Again, I presume here that developers care about providing the value to the end user. To get feedback and avoid hiccups during value delivery process developers rely on help of operations teams. And operations teams need tooling to control deployment variability, automate manual configurations, or decrease times for failure recovery. They need something like…Kubernetes.</p><p>Here I reason that introduction of such a system like Kubernetes help developers to reach their goals. And, as I mentioned above, introduction of such system has cognitive load costs to the developers</p><p>But can we actually keep the positive part and decrease negative? Can we somehow manage that mental overload of developers?</p><h4>Extraneous cognitive load</h4><p>Skelton and Pais talk about three kinds of cognitive load in their book “Team Topologies” :</p><ul><li>Intrinsic (skills)</li><li>Extraneous (mechanics)</li><li>Germane (business value)</li></ul><p>We are interested here in <em>extraneous cognitive load</em> as Kubernetes usage increases it. Extraneous cognitive load relates to the environment in which the task is being done (e.g. “How do I deploy this component, again?” “How do I configure this service?”). Skelton and Pais suggest organizations to attempt to minimize unnecessary extraneous cognitive load. By decreasing it, there is more space left for <em>germane cognitive load</em> (where “value add” thinking is).</p><h4>Digital platforms to the rescue</h4><p>Digital platforms is one popular way how companies introduce such new technologies as Kubernetes and at the same time keep cognitive teams of engineers at sustainable level.</p><p>According to Evan Bottcher, technical leader at ThoughtWorks, a digital platform could be described as</p><blockquote>“A foundation of self-service APIs, tools, services, knowledge and support, all arranged as a compelling internal product”.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/853/1*k-6AxD1Ub9N2WEqcnRFvnQ.png" /></figure><p>Digital platform aims to expose suitable level of abstractions of underlying tools and systems (e.g. Kubernetes) to a level of understanding suitable for developers of the organization. Needs of engineering teams vary in every organization and therefore such digital platforms should be tailored to those needs.</p><p>A dedicated platform team is a common way of building and managing digital platform. Product management principles should guide building it. The platform could consist of tools, automation scripts, documentation and tech support. By focusing on needs of engineers the platform team enables them to profit from technology and not to get overloaded by it.</p><p>Not having such internal platform leads to inconsistencies in the use of external services. It also leads to unreasonable demands on product teams that are already being pulled in many directions while ironically being pressured to deliver more feature faster, since they now have Kubernetes.</p><h4>Thinnest viable platform</h4><p>Digital platform size depends on the needs. It doesn’t have to be massive. Skelton and Pais recommend to always aim at “<em>thinnest viable platform</em>” to avoid the platform growth becoming an end-goal for platform team. According to Allan Kelly “Software developers love building platforms and, without strong product management input, will create bigger platform than needed”. Therefore the platform team should keep the focus on their user’s needs — the developers, and ensure the platform is built to serve those and elevate their value flow. The growth of a digital platform should be tied to the growth of technical needs the organization.</p><p>To conclude, by reducing cognitive load of developers, they are empowered to do what they do best — build and adapt features to bring value to their users. And a suitable internal platform helps to focus them on increasing delivery flow and be more effective. The key idea here is that such platform should be built for developers, their needs being primary driver of platform functionality and usage experience.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11ab68a3bad6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/swlh/how-organizations-can-add-new-tech-and-avoid-developer-burnout-11ab68a3bad6">How Organizations Can Add New Tech and Avoid Developer Burnout</a> was originally published in <a href="https://medium.com/swlh">The Startup</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Innovation Fails in Big Organizations]]></title>
            <link>https://vytotas.medium.com/why-innovation-fails-in-big-organizations-c652b119b812?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/c652b119b812</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[product-thinking]]></category>
            <category><![CDATA[product]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Thu, 11 Feb 2021 13:40:01 GMT</pubDate>
            <atom:updated>2021-02-16T14:53:02.252Z</atom:updated>
            <content:encoded><![CDATA[<p>Innovation by committee, business cases, roadmaps and projects — four horsemen of failed innovation?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*1L43c8D6fD7yoLIB" /><figcaption>Photo by <a href="https://unsplash.com/@helloimnik?utm_source=medium&amp;utm_medium=referral">Hello I&#39;m Nik 🎞</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>This blog post is a summary of a chapter from “Inspired” by Marty Cagan. I wrote this summary as reflection exercise on a topic which is critical to all Enterprises — challenges of innovation in large companies.</p><h4><strong>Innovation and Enterprise</strong></h4><p>Large enterprises leverage their services through the value and the brand created many years earlier. It is a natural thing to do. Groups of stakeholders focus on protecting what the company has created in the past. Unfortunately, this may mean putting so many obstacles to new ideas that almost nobody is willing or able to take the company to new directions.</p><p>When a company was young it likely had a compelling vision which is largely achieved now. And now the people may not be sure what’s next. Product and engineering teams complain about the lack of vision and empowerment. The fact that it takes forever to get a decision made, and the product work is turning into design by committee.</p><p>Company leadership may <em>also</em> not be happy too due the lack of innovation in the teams. They sometimes attempt to solve it by creating semi-independent “innovation centers” to facilitate new products in protected environment. Yet this rarely brings the innovation the company is desparate for.</p><p>So maybe large-scale innovation is not possible in enterprise companies? We see that it is not true. Companies such as Adobe, Amazon, Netflix continue innovate successfully even at high growth rate. The fact is that the other enterprises could do it too. But that requires making big changes.</p><h4><strong>Root Causes of Failed Products</strong></h4><p>In enterprise companies ideas for new products usually come from key company stakeholders or customers. To plan implementation of those ideas the companies want to prioritize and put them on a kind of products roadmap. They have two reasons for that. Firstly, stakeholders want product/engineering teams to work on the most valuable ideas first. Secondly, they want to know when the products will be ready.</p><p>Roadmaps are usually created in quarterly or annual planning sessions. Leaders discuss ideas and negotiate a product roadmap. And to prioritize the product roadmap, they need a <em>business case</em> for each idea.</p><p>Business cases are usually aim to answer two questions:</p><ul><li>How much money will a product make?</li><li>How much money and time it will cost?</li></ul><p>Based on the answers, the product roadmap is usually created and prioritized for 6 or 12 months ahead.</p><p>When the roadmap is confirmed, the product and technology teams get their work plans defined and communicated. Normally they will work according to that plan from the items of highest priority and down.</p><p>Once the most important idea is picked up by product manager, he usually meets the stakeholders to shape the idea and work on set of requirements.</p><p>Requirements often come as user stories. These user stories aim at explaining to engineers, designers and QA team what should be built.</p><p>Once requirements are ready, UX/UI team is tasked to provide interaction and visual designs.</p><p>When design specifications are ready, they reach engineers. The engineers use those designs and user stories to break all the work in some set of iterations. Often the idea is planned to be built as proof-of-concept (POC) or minimum-viable-product (MVP) first.</p><p>The work done by engineers and followed up by QA team. QA executes some testing to validate that the resulting application works as defined in the requirements.</p><p>This is how most of companies, large or small work. These companies, however very often aren’t satisfied with level of innovation. They also not happen that it takes a very long time to make it from the inception of an idea to a ready-to-use product.</p><p>All the steps can be illustrated in a simple diagram:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/434/1*Bug-7Per8kXNqdL74FK_uw.jpeg" /><figcaption>New product cycle as waterfall process by Marty Cagan</figcaption></figure><p>Although most of companies think they follow practices of Agile, the described process is actually <em>waterfall.</em> Waterfall process has been bashed for past 20 years, but how exactly is that a problem? Marty Cagan lists 10 reasons of how this process blocks innovation in the companies. I will dive-deep into four of them.</p><h4>1. The source of ideas</h4><p>Products developed in waterfall model products are usually stakeholder or client-driven. They focus around features facilitating sales. This may be not the source of best product ideas. Stakeholders rarely have enough time to dive deep into ideas of how products can solve particular problems of users. Another problem with stakeholder-driven innovation approach is the lack of team empowerment. The team is there just to implement. This is very similar to contractors, who focus on execution and (almost) never on innovation.</p><h4><strong>2. Business cases</strong></h4><p>Usually companies create their roadmaps based on business cases. Business stakeholders take ideas and ask two questions: how much money will it make and how much will it cost to build? Based on the answers the roadmap priorities are organized.</p><p>The problem with this approach at this stage is that nobody can really know how if a solution will be success or not. <em>Best product teams</em> <em>know </em>that half of their ideas about products and new features <em>will fail. </em>If the product team does a great job, then solution could be very successful and have an impact on a big part of the company. More often than not, however, solution ideas end up not making into successful products.</p><p>This is a key product development lesson. At this stage <em>we cannot know </em>how much money will a product make. In a same way we cannot know how much money will it cost.</p><p>The second challenge with business cases is knowing how much it will cost to build. As we don’t know how solution will look like. At this stage costs estimation is mostly educated guesses.</p><p>Summing up — business cases aim to answer two important questions: how much money will it bring and how much money will it cost to build. When an idea is at inception level, answer to these questions are guesstimates only.</p><h4><strong>3. Product roadmaps</strong></h4><p>They mostly consist of a list of products and features often prioritized according to business cases. We just seen that early business cases are guesstimates at best. Cagan mentions two insights he calls most important problem of product. Two “inconvenient truths”.</p><p>The first inconvenient truth is that not all our ideas are going to work. There could be many reasons for that. The usual one is that our customers aren’t just as exited about our idea as we are.</p><p>The second inconvenient truth with ideas is that even successful ones usually take several iterations with implementation till they start providing business value to users.</p><p>According to Cagan, there is no way to escape these inconvenient truths. Difference between successful product teams and others is how they approach these inconvenient truths.</p><h4>4. Outputs vs outcomes</h4><p>Entire process is project-centric. Companies usually fund, staff, and launch projects. But projects are output and products are about outcome. This results into many orphaned projects. Something got released, but if it doesn’t satisfy the goals, then why to keep working on it?</p><p>The biggest flaw is of such waterfall process is that we all risks remain to the end of a project. Validation with customers happen too late, when most of the budget is spent already. If the idea proved to be wrong, we can’t get our time or money back.</p><h4><strong>Summing up</strong></h4><p>Digital product innovation is hard. The way most enterprises think about the digital innovation seem to be sensible, however it doesn’t work that well.</p><p>The good news, that many companies already figured out how to make new digital products successfully. Their lessons learned are based on similar principles and ways of working. And Cagan’s “Inspired” is just about that.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c652b119b812" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Internal Platforms as Products]]></title>
            <link>https://vytotas.medium.com/internal-platforms-as-products-a8a30c17d856?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8a30c17d856</guid>
            <category><![CDATA[product-development]]></category>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[platform]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Wed, 03 Feb 2021 08:26:14 GMT</pubDate>
            <atom:updated>2021-02-11T19:42:31.029Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>What is internal platform?</strong></h3><p>The rise of DevOps movement popularized many new ideas such as <em>Continuous Integration</em> and <em>Continuous Deployment</em>, <em>Observability</em>, <em>Microservices</em>, etc. New supporting technologies like Docker and Kubernetes emerged and enabled wide adoption of those ideas.</p><p>New ways of working need new tools, technologies and environments. Tools to enable teams develop digital products faster and run them more efficiently.</p><h3><strong>Building blocks of internal platform</strong></h3><p>These tools form <strong>internal platforms</strong> which engineering teams rely on. The schematics of an internal platform could look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/842/1*Tv_garIYiKieV_LFby7faQ.jpeg" /></figure><p>The diagram shows how mutliple tools integrated together form an internal platform. Different companies may use different tools as building blocks.</p><p>To better illustrate a typical role of building blocks I will give a few examples such tools.</p><p>A good example is <a href="https://www.elastic.co/what-is/elk-stack">ELK stack </a>— a popular solution to manage logs of many applications in one place. Generally speaking, ELK consists of log database (Elasticsearch), log parser (Logstash) and User Interface (Kibana). Many plugins are available to adapt ELK to specific needs of engineering teams.</p><p>Another example is <a href="https://azure.microsoft.com/en-us/services/devops/">Azure DevOps</a> developed by Microsoft. Engineering teams use it for application planning, development, delivery, and operations. Azure Devops allows to write tasks, manage source code, build and deploy applications and much more.</p><p>These tools are flexible enough to be adopted to different needs of technical teams. There is a cost to using them too. Tools configuration requires expert knowledge of modern application development and operations principles.</p><p>Platform teams also need in-depth technical knowledge of infrastructure, automation and security topics. This knowledge is a critical part of building a well-orchestrated internal platform.</p><h3><strong>Why should we care about internal platforms?</strong></h3><p>Internal platforms are growing in popularity. Especially in enterprises that need to deliver products faster and more efficiently. The rise of <em>automation</em>, <em>big data</em>, <em>machine learning</em> adds a lot of technological complexity.</p><p>A single developer cannot be an expert anymore in every aspect of application development and operations. Internal platforms give a way to “hide” big parts of technical complexity and decrease the cognitive load of a developer.</p><p>Manuel Pais and Matthew Skelton, the authors of “Team Topologies” highlight that</p><blockquote>Developers have more and more demands on them to be able to operate, test, release their systems and run the applications in production as well, they need some kind of help, they need some support because that’s a lot of cognitive load, a lot of demand on them.</blockquote><h3>How internal platforms are build?</h3><p>Internal platforms are made from building blocks assembled to support software development pipelines. It is usually not about some particular programming language or software methodology. Companies may use ready-made blocks or may develop their own.</p><p>Typical pipeline building blocks:</p><ul><li>monitoring system</li><li>log aggregator</li><li>CI/CD</li><li>Compute orchestrator</li><li>Container orchestrator</li></ul><p>An example of an internal platform could consist of ELK stack (monitoring and log aggregation) with Azure DevOps (task management, CI/CD) and Kubernetes (container orchestration).</p><p>Large tech companies like Google or Amazon often choose to develop their own building blocks from scratch. In fact, origins of Kubernetes come from internal Google platform called “Borg”. The company required enormous infrastructure to make its search engine available to billions of users.</p><p>But not all the companies have needs and resources of Google or Amazon. Platform engineers should not try to build everything from scratch. Instead they should rely on using existing commercial or open-source solutions. Focus should be empowering technical teams instead of building tools for problems which are solved already.</p><p>Taking ready-made platforms and making them work seamlessly together is not an easy task. It requires writing code. Platform team needs to understand how applications are currently built, developed and operated. And how these processes can be supported and improved by an internal platform.</p><h3>Why internal platforms should be built as products?</h3><p>Internal platforms usually are build differently than digital products. Internal platform projects often start by small teams of tech-savvy engineers. They often are experts on multiple technologies and follow latest industry trends and best practices. Such platform teams focus on technology and technical integration most of their time.</p><p>What platform teams often miss is focus on future platform users. The users are expected to leverage on the internal platform to do their work. When platform team does not pay enough attention to the needs of users they risk to build something too complex to use. Such platform may also solve problems that users don’t have. And lack of positive user feedback discourages the platform team to finalize their work.</p><p>A more effective way to build internal platforms is to treat them in the same way as an external product offering. Applying product management approach to internal platforms means building empathy with internal users and working with them on the design.</p><p>Product management approach would call for a product development strategy. Big part of this strategy is to understand what are the outcomes and impact to be achieved. To achieve greatest impact a platform team must pay very close attention to the gaps and pain points engineering teams have in their application delivery worklow.</p><h3>Why is building internal platforms as products difficult?</h3><p>Building internal platforms as product seems to be a good way to maximize the level of adoption of such platforms. But what are the challenges of applying such strategy?</p><p>In the popular post <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=&amp;cad=rja&amp;uact=8&amp;ved=2ahUKEwjX7rG-ks3uAhUH6qQKHe1TCDwQFjAAegQIAhAC&amp;url=https%3A%2F%2Fskamille.medium.com%2Fproduct-for-internal-platforms-9205c3a08142&amp;usg=AOvVaw0zAfMiuJoPQYNrE_gwR7KF">“Product for Internal Platforms”</a> <a href="https://medium.com/u/12421a4f856">Camille Fournier</a> highlights a few reasons why this happens:</p><ul><li><em>Customer group size is small</em>. It is difficult to build using standard metrics-based product strategy. When you do not have thousands of users the usual A/B testing will not tell you much.</li><li><em>Captive audience</em> . If your users do not have a choice but have to use your tool then of not using your tool then adoption metrics aren’t that useful. Platform engineers are less motivated to develop fully finished product too.</li><li><em>Hard to think like your customer</em>. Platform developers are experts of the platform. It is difficult for them to understand how it is to use their platform for someone who uses it occassionally. And is not expert of every aspect of application lifecycle.</li></ul><h3>Summary</h3><p>Platform teams should know how products they build make the overall engineering team more effective. Focus on a value for the users is something that often missed in internal platform projects.</p><p>A clear strategy is needed not only for internal platform development process. Application migration process, user onboarding, platform evolution aspects need clear strategies as well.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8a30c17d856" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ways to Modernize Legacy Application]]></title>
            <link>https://vytotas.medium.com/ways-to-modernize-legacy-application-d0f6d66a999c?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/d0f6d66a999c</guid>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Tue, 12 Jan 2021 07:48:06 GMT</pubDate>
            <atom:updated>2021-02-03T21:02:37.589Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*oueGTU31vsAcKBFa.jpeg" /></figure><p>In my previous post “<a href="https://vytotas.medium.com/legacy-applications-costs-e9aafebd3b0d">Legacy Application Costs</a>” I listed common costs and risks of legacy applications. In this post I will give an overview of different ways of modernizing those applications.</p><p>Every now and then engineering teams face situation when they need to decide what to do with a legacy application. Questions they often ask themselves are:</p><ul><li>What if we keep the application as it is?</li><li>How should we modernize the application? Should we refactor it? Should we rewrite it from scratch?</li><li>What is the total application cost of ownership now? Are we able to estimate it? What will be the cost of ownership after the modernization?</li><li>Can we complete modernization project under planned time and budget? Do we have people with the knowledge and skills to make this modernization?</li></ul><p>Decisions of how to modernize applications are rarely easy. High level of technical complexity, high modernization costs, business and technical risks are just a few challenges which need to be assessed. These challenges can also hinder engineering teams from being able to see the complete picture.</p><p>What engineering teams often do not pay enough attention to is the negative impact legacy application has on business. High costs<strong> </strong>of<strong> </strong>maintenance over time, low organizational efficiency and maybe even lost business opportunities are just few examples. Understanding this impact, however, helps when talking to business stakeholders.</p><p>Measuring legacy application impact on business is often difficult. It requires organizational maturity to define needed metrics correctly. It requires even more maturity to continuously collect measurements through longer periods of time. And some aspects aren’t just not quantifiable.</p><h3>Business value as critical success factor</h3><p><a href="https://www.infoq.com/articles/10-tips-prevent-business-value-risk/">According to</a> Chris Matts, there are only 4 reasons to invest into IT from a business perspective:</p><ul><li>Increase revenue</li><li>Decrease costs</li><li>Protect revenue</li><li>Avoid costs</li></ul><p>This concept can also be applied to application modernization projects.</p><p>For example, are we aiming to <em>reduce costs of application maintenance</em>? Then we need to determine what exactly makes the maintenance expensive. Maybe it is inflexible codebase, maybe outdated infrastructure. The cost driver of maintenance should drive decision of what modernization should include and, often more importantly, what not.</p><p>During the whole modernization process having clarity on business goals will be a key factor to achieve the expected business impact. Resuming the previous example about application maintenance cost reduction :</p><ul><li>How to ensure that modernization substantially reduced maintenance costs?</li><li>How to keep focus on the work which impacts cost? Engineers love making application code cleaner or architecture more scalable in modernization work. However these important attributes of applications may not significantly impact the cost.</li><li>How to involve business stakeholders in prioritization of work so that most valuable items for cost reduction are worked on first?</li></ul><h3><strong>Choosing modernization technique</strong></h3><p>According to <a href="https://www.gartner.com/smarterwithgartner/7-options-to-modernize-legacy-systems/">Gartner research</a>, the problems of legacy applications usually come from either <em>technology</em>, <em>architecture</em> or <em>functionality</em>.</p><p><em>Technology</em> and <em>architecture</em> issues may cause application costs to be too high, the code too complex or create other risks (security, scalability, etc. risks).</p><p>If application <em>functionality</em> does not meet current business needs, then revenue growth or organizational agility opportunities may be lost.</p><p>A few examples:</p><p><strong>Technology</strong></p><ul><li>The underlying technology is not supported by vendors anymore</li><li>The technology licensing is too expensive</li></ul><p><strong>Architecture</strong></p><ul><li>The application architecture maintenance costs are too high</li><li>The application cannot be scaled to meet the business needs</li></ul><p><strong>Application functionality</strong></p><ul><li>The new features needed by business users cannot be added</li></ul><p>Identification of a problem source being technology, architecture or functionality helps further with a choice of application modernization technique.</p><p>Personally it was helpful to analyze the legacy applications and think if the problem comes from the technology (that framework we didn’t update for several years) or the architecture (that other app with spaghetti code) or the functionality (we just cannot have automated processing of PDFs the customers keep asking).</p><p>Of course in some cases it was a mix of two or all three problem sources. Nevertheless the exercise of thinking and discussing proved to be very valuable for a better understanding.</p><p>Gartner researchers also recommend to choose legacy application modernization technique based on problem source. The diagram below illustrates relationship between the modernization technique, the implementation effort and the expected impact. The diagram also helps to limit a number of suitable modernization techniques and put the focus on the most relevant ones.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*luF7puD8T0NbQMK4.png" /></figure><h4>What are the legacy modernization techniques?</h4><p>There are seven modernization techniques in the diagram. The most known ones in the IT industry are <em>refactoring</em> and <em>rebuilding</em>. It is worth to look at all of them so that you can consider the best ratio between risk and reward.</p><ul><li><strong>Encapsulation</strong> — creating wrapping around a legacy application or its component that provides a new interface making it easily accessible to other software components. Small changes to the code make the risks minimal.</li><li><strong>Re-hosting</strong> — moving application unchanged to other (usually cloud) infrastructure. This technique is of the lowest cost and risk.</li><li><strong>Platform migration</strong> — adjusting the code to a new platform while preserving existing functionality. The minimal changes like using a managed database or adding auto-scaling.</li><li><strong>Code refactoring </strong>— restructuring and optimizing application code without changing its external behavior.</li><li><strong>Re-architecting</strong> — shifting to a new application architecture while altering the code to use the new capabilities of the platform.</li><li><strong>Replacing</strong> — entirely replacing the app with a different tool rather than investing in its modernization.</li></ul><h3>Conclusion</h3><p>The business value of modernization needs to be identified and shared by application stakeholders. Having clear business goals will help to define modernization scope and measure success.</p><p>By looking at the legacy modernization from different perspectives we decrease a risk of overpaying while still not solving the core issue.</p><p>Technology, application architecture and functionality are the three main causes of why legacy problems occur. The mapping between the causes and the modernization techniques helps to limit the choice and also balance between modernization effort and expected impact.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d0f6d66a999c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Legacy applications costs]]></title>
            <link>https://vytotas.medium.com/legacy-applications-costs-e9aafebd3b0d?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/e9aafebd3b0d</guid>
            <category><![CDATA[refactoring]]></category>
            <category><![CDATA[legacy-code]]></category>
            <category><![CDATA[total-cost-of-ownership]]></category>
            <category><![CDATA[application-development]]></category>
            <category><![CDATA[legacy-systems]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Fri, 01 Jan 2021 10:41:34 GMT</pubDate>
            <atom:updated>2021-02-07T10:20:07.252Z</atom:updated>
            <content:encoded><![CDATA[<h3>Legacy application costs</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/960/1*VWCXe8GOx4LnHjTRSXwwHg.jpeg" /></figure><h3>Introduction</h3><p>Applications based on legacy code exist in most mid-to-large companies. At the company I work we have them too. It is not that difficult to identify those applications. Development teams know them well as they usually require a lot of attention. Yet a deeper look reveals the reasons why we consider those application legacy are different. And they pose different downsides as well as risks.</p><p>My motivation for this post is to look at common attributes of legacy applications. What are the costs and risks of not addressing the legacy aspect of those applications?</p><h4>What is legacy code?</h4><p>There is a plethora of definitions of legacy code. Below are a few of the most popular:</p><ul><li>Code which uses no longer supported technology</li><li>Code without tests</li><li>Code written by another team than maintaining one</li><li>Code that has different conventions than ones used by maintainers</li></ul><h4><strong>Who should care about the application being a legacy?</strong></h4><p>The most vocal people about legacy applications are developers. Legacy code impacts developers in daily work. Yet, legacy applications impact business as well. Legacy can pose business risks and incur significant costs. I identified 4 groups of people affected by code being legacy:</p><ul><li>Developers</li><li>Engineering managers (responsible for organizing work to keep the application running)</li><li>Clients of the application (usually business people who bear application maintenance costs)</li><li>Users of the application (business people who use the application to do their work)</li></ul><h4><strong>Why should they care about code being legacy?</strong></h4><p><strong>Changes take a longer time</strong></p><p>Making changes in legacy code is hard. Changes take much more time compared to well-maintained code. Lack of documentation, poor code structure, no tests, and tight coupling make changes risky. Developers take extra time to gain confidence that nothing gets broken by accident.</p><p><strong>Changes causing unexpected application behavior</strong></p><p>Changes in legacy code may cause an application to start having unexpected behavior. We call it regression. Developers change the code to add a feature. That code change introduces unwanted application behavior. Sometimes weeks pass before someone notices that the application has bugs. If more changes were added meanwhile, then it is also difficult to track back which change caused it.</p><p>We use testing to catch unexpected application behavior when we add some new feature or fix a bug. Legacy projects often lack automated tests. They also lack documentation for testing the application behavior manually.</p><p><strong>Application scaling difficulties</strong></p><p>Applications used for a long time may need more resources than originally planned. Scaling the infrastructure of a legacy application is usually very challenging.</p><p>Imagine a situation where a legacy application consumes all available server memory. We could try to split the application between several servers to gain more resources. If the application architecture is not modular then such a split is not possible. The application needs refactoring first.</p><p><strong>Legacy project staffing</strong></p><p>Developers usually aren’t motivated to work on legacy code. It takes them a lot of time to understand why an application works the way it works. It may take weeks of reading badly-structured code and reverse-engineering it. And writing code is always easier than reading someone else’s.</p><p>Developers also often fear to add bigger changes to legacy applications. There is always a risk of bigger changes introducing unwanted behavior.</p><p>Legacy applications often rely on outdated and no longer supported technologies. A company would need to find developers who know such technologies and it doesn’t come cheap. Developers also not keen to learn such outdated technologies. Most of them prefer to learn skills useful in the future. Skills that open opportunities to join new projects.</p><p>Developers may also fear to be stuck maintaining projects based on outdated technologies. It is much easier for a company to staff new projects than legacy projects.</p><h4><strong>What are the hidden costs for business?</strong></h4><p>Long lifecycle for changes, unexpected application behavior and staffing challenges are the problems that affect development teams. The negative effects of legacy incur additional costs to business as well. These costs aren’t directly visible and often cannot be quantified easily.</p><p><strong>Maintenance and Support</strong></p><p>Maintenance cost can exceed the original development budget just 3-5 years after release on an application. <a href="https://pdfs.semanticscholar.org/7eee/629b22cd3db63296cac13a0c37cb0a7235f6.pdf">Research</a> shows that on average the maintenance consumes around 60% of software costs.</p><p><strong>Integration and Compliance</strong></p><p>Failure to comply with industry regulations could potentially lead to penalties. Recent changes in the GDPR area made it a real concern for most companies and not just ones specialized in regulated industries.</p><p><strong>Security</strong></p><p>Legacy code can lead to data breaches, which can lead to financial and reputational losses. Cybercriminals are armed with the knowledge of the weaknesses in older systems and those older systems are used to run legacy code.</p><p><strong>Lost Business Opportunities</strong></p><p>The lack of innovation and development speed can impact profitability and competitiveness. Not being able to meet new customer needs under reasonable time and budget may cause a company to lose business opportunities. In long-term, a company could lose its market share given legacy application is critical to the business.</p><p><strong>Organizational Agility and Efficiency</strong></p><p>The average efficiency of some processes can increase multifold through automation and modernization. Many manual steps performed by organizations can be either automated or eliminated at all. Think of all those steps business teams perform daily to clean and prepare data. Or the amount of hours those teams need to spend to test a legacy application after some minor change.</p><h4>Conclusion</h4><p>I conclude that legacy applications have hidden costs which we often oversee. It was helpful to list those costs as I can use it to analyze individual applications further.</p><p>Identification of costs and risks is the first step in managing legacy applications. In my next post, I will explore different legacy application modernization techniques.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e9aafebd3b0d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Challenges of setting OKRs up]]></title>
            <link>https://vytotas.medium.com/challenges-of-setting-okrs-up-d62d8d94e414?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/d62d8d94e414</guid>
            <category><![CDATA[okr]]></category>
            <category><![CDATA[objectives-key-results]]></category>
            <category><![CDATA[goals-objectives]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Sat, 01 Feb 2020 12:32:44 GMT</pubDate>
            <atom:updated>2020-02-01T12:32:44.245Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MWlCMiPhY99_E1VD2UEmCg.jpeg" /></figure><h3><strong>What are OKRs and why should we care?</strong></h3><p>OKRs stand for “<em>Objectives, Key Results” </em>and it is a goal setting system for organizations. OKRs is one of the most popular systems in IT world right now. Some other systems which were popular in the past and still used now to set goals are KPIs, Management by Objectives, SMART goals.</p><p>The purpose of OKRs is to connect the goals of individuals, teams and the whole organization to measurable results while helping all the members of the organization to move in the same direction.</p><p>In my opinion three things make OKRs different from other popular systems.</p><p>OKRs emphasize importance of making objectives outcome-oriented (opposite to output-oriented). In other words, focusing on results rather than on activities.</p><p>OKRs provide possibility to share same objectives across multiple teams. Shared objectives help multiple teams contribute to the same outcome-oriented goals.</p><p>OKRs stress the importance of not being used for performance evaluation of individuals. By not being a tool to impact salaries and bonuses, the process of setting objectives and key results becomes “gamed” by individuals much less.</p><p>Internet is full of articles describing how to set up OKRs. Instead of repeating the same story I am going to focus on common challenges of implementing the system in an organization.</p><h3><strong>Challenges in using OKRs</strong></h3><p>I see three groups of challenges when implementing or using OKRs. The first group relates to defining OKRs, the second group is about challenges of having OKRs shared across the organization. The third group relates to general problems of OKRs as a goal setting system.</p><p><strong>Defining OKRs</strong></p><ul><li><em>Measuring results</em>. One of the rules of OKRs is that key results should be defined in a way that their progress could be measured easily. Formulating a key result as “Improve X by 30%” may not be concrete enough. What you will you measure exactly and what will you compare it to? At what moment will you do the comparison?</li><li><em>Isolating results which are affected by many teams</em>. Engineering team’s “decrease average time-to-market for features by 30%” objective can be affected by Operations team’s “decrease deployment time by 20%” objective. Therefore Engineering’s team’s objective may be impacted even if the team does not work on it.</li><li><em>Mixing top-down and bottom-up OKRs</em>. There is a risk for all objectives and key results of a team to be set by management. It may have negative effects on the team’s motivation and ownership of key results. In addition, team may have very good ideas about objectives and KRs too since it is closer to actual work. <br>A healthy mix of top-down and bottom-up OKRs helps to achieve company-level alignment and relevance of OKRs to actual needs of teams.</li><li><em>Solution-specific vs. business-outcome specific OKRs. </em>A key to implementing OKR system successfully is to formulate them with a business outcome in mind. The power of OKRs is in focus on the results and not activities. “<em>Implement automated system for sending emails</em>” is an activity whereas “Cut the time needed to send emails by 80%” is an outcome.</li><li><em>Right granularity to enable team synergies. </em>OKRs which have too broad granularity can make teams loose focus and miss opportunities to create synergies. A common objective ‘improve time-to-market by 20%’ for Engineering and IT Operations teams leave many paths to make an impact independently from each other. However “improve time-to-market by 30% by through releasing 50% less defects” objective let the teams create synergy by focusing on deployment and automated testing area together.</li><li><em>Too many OKRs can backfire as the team may loose focus. </em>OKRs help to move companies with a focus towards same direction. Too many OKRs can confuse teams and make them loose that focus. General recommendation is to start form 3–5 objectives.</li></ul><p><strong>Sharing OKRs with other teams</strong></p><ul><li><em>OKRs driven from one team to another</em>. OKRs help create shared objectives across the organization. This means that some teams will request other teams to commit on certain OKRs. Should the team prioritize their own objectives over the objectives requested by another team? Probably the only way to answer that is to look at the high-level objectives of the company and then decide which objectives the receiving team should pick.</li><li><em>Hierarchical OKRs</em>. Some organizations choose to structure their OKRs in a hierarchical fashion for all teams. This means that Key Results of a CEO turn into Objectives of departments. Key Results of departments become Objectives of department teams.<br>Other organizations think that effort needed to align OKRs into such hierarchical manner is too big. As an alternative it is then suggested for teams to look at organization-level OKRs and define teams OKRs so that they positively impact one or more organizational OKRs.</li></ul><p><strong>General challenges</strong></p><ul><li>Complexity in setting OKRs, requires training. OKR approach seems deceivingly simple. However the process of OKR implementation may be difficult. In a scenario where the whole company attempts to define their objectives for the first time, the organizational effort is very high.</li><li>OKRs do not cover performance evaluations. One of key success factors for OKR implementation is not using it for employee performance evaluations. The reason is simple — to keep teams setting ambitious goals and not “sand-bagged” ones. This means there should be a separate process for performance evaluation.</li><li>If KRs assigned to a team, there may be accountability issues. Many organizations which use OKRs prefer not to have OKRs on individual level, but stop at the level of a team. Such choice seems to be valuable when the organization depends a lot on collaboration in teams. In such case accountability issues may arise when nobody in a team feels accountable for achieving objectives, because they aren’t individual</li></ul><p><strong>Conclusion</strong></p><p>In this article I discussed numerous challenges organization can face when implementing OKRs. Challenges are numerous and it may take a few quarters for an organization to find an OKR setup which works. However, not having some sort of a system which helps the organization to align across all teams may have negative consequences in a long run, especially within large organizations.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d62d8d94e414" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Understanding Dreyfus Model]]></title>
            <link>https://vytotas.medium.com/understanding-dreyfus-model-a58762ac3e81?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/a58762ac3e81</guid>
            <category><![CDATA[education]]></category>
            <category><![CDATA[dreyfus]]></category>
            <category><![CDATA[skills-development]]></category>
            <category><![CDATA[learn]]></category>
            <category><![CDATA[learning]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Sat, 11 Jan 2020 20:22:04 GMT</pubDate>
            <atom:updated>2021-01-17T16:26:40.454Z</atom:updated>
            <content:encoded><![CDATA[<h3>How We Learn: Dreyfus Model</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*XqvZXIMCxaob-BdWNV2AOw.png" /></figure><h3><strong>Introduction</strong></h3><p>One of the key differences between our professional lives now and 20 years ago is a concept of life-long learning. Importance of continuing learning all the time seem to be a prerequisite many more professions now than in the 20th century.</p><p>To be effective in this learning journey we need to understand the stages of acquiring skills. What makes a difference between novice and master? Is this only practical knowledge? Or do they see problems and tasks differently? Why a master doing the task looks so fluent whereas a novice is under great stress with exact same task?</p><p>In this article I am going to discuss Dreyfus model which explains different stages of skills one progressively gains. I will describe the stages from novice to master. I will also explain three key dimensions of our mental processes which change due when we progress in our skill.</p><h3><strong>Stages</strong></h3><p>Skill acquisition topic has been popular among researchers for quite a while already. One of the most famous models in this area is Dreyfus skill acquisition model. It was formulated in 80s based on research on how pilots, nurses, second language learners and chess players progress in their skills. Although the model is 50 years old, it still is relevant today.</p><p>Dreyfus model identifies five stages in skill acquisition: novice, advanced beginner, competent, proficient and expert.</p><h4><strong><em>Novice</em></strong></h4><p>Novices do not particularly wants to learn, they just want to get the job done. Novices cannot identify what parts of a problem are important and which aren’t. He can only use rules if they are context-free, such as “in case of A do B”. They do not know how to deal with mistakes.</p><h4><strong><em>Advanced Beginner</em></strong></h4><p>Advanced beginners start breaking rules, but just barely. They can use some advice and apply it to familiar context, but to the limited extent. Advanced beginners want information fast, however they aren’t interested in a bigger picture. If you try to explain the bigger picture they will dismiss is it as irrelevant. This is still “looking at trees, not seeing a forest”.</p><h4><strong><em>Competent</em></strong></h4><p>Competent start using forming conceptual models of a problem domain. They deliberately plan their work based on their past experience. Competent can troubleshoot on their own. They can apply expert’s advice. When lacking past experience, they still have trouble in determining which details are important and should be focused on. They cannot self-correct on this stage.</p><h4><strong><em>Proficient</em></strong></h4><p>Proficient need and understand big picture. If information is over-simplified it brings frustration to them. Proficient can learn from past mistakes and are able to improve their performance based on past experience by deciding what to to differently next time. Proficient can learn from experience of others and case studies.</p><p>Proficient can understand maxims, which are fundamental trues and apply them correctly. For example, “test everything that can break” maxim will not bring value to novice, who may decide to test every smaller part of a software application. Proficient, however, will be able to reflect on what typically breaks in software applications and device proper testing strategy.</p><p>Proficient know what is coming next based on the past experience. If it doesn’t happen, they know what need to be changed.</p><p>Proficient take full advantage of reflection and feedback.</p><h4><strong><em>Expert</em></strong></h4><p>Experts have vast knowledge on the subject and can apply right part of it for a particular context. Instead of following rules, they base their decisions on the intuition. When asked to explain what rules they used to solve some particular problem, experts often cannot. They acted in some way because “it felt right”.</p><h3><strong>Dimensions</strong></h3><p>You probably recognized already several common dimensions being discussed in each of the stages. It is important to mention them as they are helpful when trying to identify a stage person is at for particular skill.</p><p>Best way to illustrated this with a diagram:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/342/1*zfJqs1TETKH26z79m2Zqig.png" /></figure><p>The first dimension is about moving from reliance on rules to intuition. Novice uses his analytical thinking and rules to make decisions. On the following stages he relies less and less on analytics and more on intuition which comes from experience.</p><p>The second dimension is about a change in perception, where a problem is not a collection of equally important bits, but a unique system of bits where some are more important than others.</p><p>The third dimension is about a change from being detached observer to being a part of involved system itself.</p><h3><strong>Conclusion</strong></h3><p>This article gave an overview of Dreyfus Model — one of the most popular skill acquisition models today. We saw that on the journey from novice to master a learner experiences the transformation in level of his intuition, ability to focus on important parts and get emotionally involved in the process and outcome.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a58762ac3e81" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Project risks from perspective of business investor]]></title>
            <link>https://vytotas.medium.com/project-risks-from-perspective-of-business-investor-6f7cc4e78fdb?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/6f7cc4e78fdb</guid>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[risk-management]]></category>
            <category><![CDATA[it-project-management]]></category>
            <category><![CDATA[business-value]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Wed, 08 Jan 2020 20:07:01 GMT</pubDate>
            <atom:updated>2020-01-08T20:09:18.148Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/970/1*voChxIdD2ZlSc-JpSumQkA.jpeg" /></figure><h3><strong>Introduction</strong></h3><p>Risk management is important topic when one considers investments into IT projects. In most cases most of attention put to project delivery risks whereas there are two other major risks from the perspective of business investment.</p><p>In this article I will give a brief overview of what are expressions of business value. I will also outline what are three major risks and give a more detailed description to one of them.</p><h3><strong>Expressions of business value</strong></h3><p>Chris Matts has identified 4 expressions of business value from business investor view why he could invest into IT project:</p><ul><li>Increase revenue</li><li>Decrease costs</li><li>Protect revenue</li><li>Avoid costs</li></ul><p>If most people are familiar with business value coming from revenue increase of costs decrease, then revenue protection and costs avoidance may need some clarification.</p><p>Protecting revenue may come from IT project which aims to develop some functionality which keeps existing customers from going to the competitor’s solution. In this case business investor makes investment to continue having same revenue.</p><p>Avoiding costs could be achieved by meeting a new regulation. A good example is GDPR regulation which may cause substantial fines given that the IT solution does not meet it conditions.</p><p>Any project which is considered to bring business value should fall into one or more of those categories. If a value of future or ongoing IT project cannot be allocated to at least one of those expressions then the red flag should be raised as the project most likely going to fail.</p><h3><strong>IT risks from investor’s point of view</strong></h3><p>There are three main risks for business investor in IT projects:</p><ul><li>Delivery risk</li><li>Business Value Failure risk</li><li>Existing Business Model risk</li></ul><p><em>Delivery risk</em> is all about project not being delivered on time, budget or quality. This is most well-known risk with many well-known techniques of mitigation.</p><p><em>Business Value Failure risk</em> is a risk of project failing to deliver expected business value. It can further be broken into uncertain valuation and communication issues risks. These risks will be discussed further in this article in more detail.</p><p><em>Existing model risk</em> is a risk of the project damaging existing business model. Good example is a project aimed at automating work of a consultant. Let’s imagine a situation when consultant spends several days with a client to install proprietary software. The client pays for two days of work of the consultant. If that IT project delivers automated way to integrate the software in an hour, the company will most probably will loose the revenue.</p><h3><strong>Business Value Failure risk</strong></h3><p>Business Value Failure risk can be split in two other risks: valuation uncertainty and communication issues. Both risks can be split further down:</p><p>1. Valuation Uncertainty:</p><ul><li>Project has no value</li><li>Project value is unknown</li><li>Project doesn’t deliver expected value</li></ul><p>2. Communication Issues</p><ul><li>Project delivers wrong business value</li><li>IT sponsor is not involved enough in the process of investing</li></ul><p><em>Project has not value. </em>Some projects are initiated without any real business value behind. If you cannot see how project delivers at least on of for value types mentioned above, then most probably there is no business value in such project. It often maybe management “ego” app. Other way to spot a project without a business value is to see if a business case talks about a solution in stead of value to be captured.</p><p><em>Project value is unknown. </em>It may happen that project has business value, but project team does not know it.</p><p><em>Project doesn’t deliver expected value. </em>This is by far most common risk. Business case description promises great revenue and when the project is delivered, it captures just a fraction of it.</p><p><em>Project delivers wrong business value. </em>Sometimes project team does not know the business value and they invent one. As a result the project is being built for different purpose than investor intended. This may especially happen with projects aimed at protecting revenue or avoiding costs.</p><p><em>IT sponsor is not involved enough in the process of investing. </em>Many IT projects has a group of departments as investors. Project team usually communicates to each of them directly. Priorities between the needs of all the departments often are done by project team alone, instead of prioritizing with the organization as a single investor.</p><p>This often happens when a project team makes the prioritization meetings painful and long. Business people tend to start avoiding those meetings leaving the project team to prioritize on its own.</p><h3><strong>Conclusion</strong></h3><p>This article has introduced four expressions of business value which I find very helpful when thinking about value of products and projects. It also gave an overview of three main risks a business investor faces when investing into IT projects.</p><p>The knowledge of types of risks and types of business values may be very helpful when assessing business cases and existing IT projects.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6f7cc4e78fdb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Observability — part 1]]></title>
            <link>https://vytotas.medium.com/observability-part-1-ef82c641c679?source=rss-172fc8740d10------2</link>
            <guid isPermaLink="false">https://medium.com/p/ef82c641c679</guid>
            <category><![CDATA[monitoring]]></category>
            <category><![CDATA[observability]]></category>
            <category><![CDATA[devops]]></category>
            <dc:creator><![CDATA[Vytas Prazarkevicius]]></dc:creator>
            <pubDate>Mon, 06 Jan 2020 22:11:09 GMT</pubDate>
            <atom:updated>2020-01-06T22:11:09.131Z</atom:updated>
            <content:encoded><![CDATA[<h3>Observability — part 1</h3><h3><strong>Introduction</strong></h3><p>Today observability is a popular topic for discussions in some technology circles. It is often viewed as a glorified monitoring of applications which is achieved by using such tools as ELK. It is, however not a case. Observability is a property of the system (and not a process) which is important for monitoring and diagnostics.</p><p>In this article I will provide a definition of observability (shamelessly stolen from VividCortex). I will also discuss how observability of a system can be achieved using two simple methods.</p><h3><strong>How do we observe systems</strong></h3><p>Observability is a property of a system which describes how well can we determine what is happening inside of the system by only observing changes in the external behavior which we observe in system monitoring data.</p><p>Monitoring is a process of checking if a system behaves as expected by observing its external behavior. Such behavior is usually expressed by telemetry (data emitted by instrumentation).</p><p>Diagnostics is a process of trying to determine why the system is not behaving as expected. This is where observability comes to play — we need the system to be observable to perform diagnostics with reasonable effort.</p><p>In order to have a good observability one often needs to build a system with to build a system so that it can be monitored and diagnosed properly. And that requires understanding what are important parts to be monitored. There are hundreds of metrics one can use, however most of them will not be useful to understand properly if a system is failing and why this is happening.</p><h3><strong>Methods for choosing metrics</strong></h3><p>One of the ways to do so is to use a simple framework, which defines two main dimensions: external quality of service and internal sufficiency of resources.</p><p>External quality of service puts a user at a center. From perspective of a user the quality of the system is good if he receives correct and fast answers to his requests. Therefore by observing how requests are answered we can determine the quality of service.</p><p>Users usually generate many requests and we should aggregate them. One popular way to look at user requests is called RED method. RED stands for rate, errors and duration (of requests). These three metrics typically can quickly answer if system is providing quality service to users.</p><p>RED metrics only tell us if we have a problem. They do not tell us why. In order to answer this question we can apply USE method which focuses of internal sufficiency of resources. The insufficiency of CPU, memory, storage or network resources are the main reason of low quality of service.</p><p>USE stands for utilization, saturation and errors. Utilization describes how of all resource capacity is used. Saturation indicates the job queue formed in front of a resource. And errors are a rate of errors coming from a resource.</p><p>In the event-based system we usually look at changes of RED and USE metrics. These changes often indicate that quality of service may be dropping. With timely action taken we can react in time to prevent the negative impact to the service.</p><h3><strong>Conclusion</strong></h3><p>Monitoring and diagnosing systems in production becomes a key activity in keeping applications running. Simple applications may be highly observable and easy to monitor and diagnose “out of the box”. In many cases, however, developers need to build applications in a way that monitoring and diagnostics is efficient. This is a trade-off we need to during application development.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ef82c641c679" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>