<?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 Chloe (Rowan) C. on Medium]]></title>
        <description><![CDATA[Stories by Chloe (Rowan) C. on Medium]]></description>
        <link>https://medium.com/@sweetpavement?source=rss-2e461d5f22c2------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*lg-hwf0KMNCUKrsY_JCXDQ.jpeg</url>
            <title>Stories by Chloe (Rowan) C. on Medium</title>
            <link>https://medium.com/@sweetpavement?source=rss-2e461d5f22c2------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 13:55:29 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@sweetpavement/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[A Night of Monitoring and Observability with BuzzFeed and DataDog!]]></title>
            <link>https://medium.com/buzzfeed-tech/a-night-of-monitoring-and-observability-with-buzzfeed-and-datadog-a7e97b10424e?source=rss-2e461d5f22c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/a7e97b10424e</guid>
            <category><![CDATA[devops]]></category>
            <category><![CDATA[monitoring]]></category>
            <category><![CDATA[tech-talk]]></category>
            <category><![CDATA[observability]]></category>
            <category><![CDATA[buzzfeed-events]]></category>
            <dc:creator><![CDATA[Chloe (Rowan) C.]]></dc:creator>
            <pubDate>Wed, 02 May 2018 17:21:52 GMT</pubDate>
            <atom:updated>2018-05-02T17:23:51.603Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gFUAvfkMzpI5dias2CBA6A.png" /></figure><p>As you can probably imagine, we have a lot of things to monitor and observe at BuzzFeed. From how long it takes for an article to load to how much CPU that loading is asking for from the server, we watch it all with help from <a href="https://www.datadoghq.com/">DataDog</a>, who live up to their tagline “modern monitoring and analytics.” Since we’re so into what they’re doing, and since we make such good use of it, we asked them to co-host an event with us about, you guessed it, monitoring! So on April 5, 2018, we threw open the doors with them for this event.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TxsGLwMG8qL6k9Pdvk4smw.png" /></figure><p>First up was <strong>Mark Hintz</strong>, a software engineer working in Data Visualization for DataDog, to talk to us about how DataDog “dogfoods” their own product by using it internally to dashboard. They don’t just monitor their own systems: they monitor some of the systems that surround them, such as the MTA (check out <a href="http://www.mtaservicechecker.com/">MTAServiceChecker</a> to see if your train is delayed!) or a marathon being run by a coworker. My personal favorite external system checker is <a href="http://ispokemongodownornot.com/">Is Pokémon Go Down or Not</a>, which tells me whether I need to bother putting on sneakers to go “catch ’em all!” Mark’s talk really set the stage for the other talks in the evening, by walking us through how DataDog handles some of the challenges that come their way: ensuring redundancy and backups, making sure that scaling capacity is seamless, and making sure that if something breaks…solving it is as close to a single button push as possible.</p><p>After Mark, we all took a quick breath before <strong>Kenton Jacobson</strong>, Director of Engineering for Vogue, Glamour, and GQ at Condé Nast, shared with us how low-noise, actionable metrics are difficult to achieve but invaluable. Using graphs from Condé Nast, including one showing the incredible traffic spikes around the Met Gala (confession, my browser is one of those data points 😉), we were able to understand how the work he had done around refining metrics really helped to drive engineers to the solutions to their problems.</p><p>After the guests, BuzzFeed took the stage, with a series of three lightning talks about monitoring and observability at BuzzFeed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/432/1*XTNJl_7t9ggnw1ggoMhSig.jpeg" /></figure><p><strong>Felicia Cippoletti,</strong> a member of the SRE team focused on databases, spoke about being familiar with day to day trends in our metrics. While having metrics is great, if we’re not familiar with what a “green” or “up” state looks like, it can be more difficult to track down the source of problems. Felicia discussed the importance of knowing what your metrics “should” look like, so you know when they’re telling you something’s wrong</p><p>After Felicia, we were treated to a bit of a “war story” from Dan Meruelo. Dan shared a time when some of our tagged metrics went missing at deploy time. He walked us through how he discovered that a code interaction between our in-house code and the DataDog metrics agent conflicted, thus leading us to think we had problems. After that dive in, he showed us what the metrics should look like and explained how to get there.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/360/1*xsp36w6VKJyIvAvr_SeasA.jpeg" /></figure><p>To close out the talks, I <a href="https://www.youtube.com/watch?v=Wd0xVvMFIUY">spoke about the four principles I used</a> in developing a proposal for how to move forward with observability at BuzzFeed. I particularly wanted to share how we were using human-focused ideas such as: reducing the burden of on-call shifts, empowering service owners who aren’t familiar with the ops domain to drive the decisions we make about code, and about how to observe our systems.</p><p>In addition to all the great information shared, it was a fantastic night where we got to open our doors to the community. Everyone got to meet new folks as well as spending time with friends and colleagues. The crew from DataDog was fantastic, and so was everyone who joined us to learn about observability. We’re already looking forward to the next one!</p><p><em>To keep in touch with us here and find out what’s going on at BuzzFeed Tech, be sure to follow us on </em><a href="http://twitter.com/buzzfeedexp"><em>Twitter @BuzzFeedExp</em></a><em> where a member of our Tech team takes over the handle for a week!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a7e97b10424e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/buzzfeed-tech/a-night-of-monitoring-and-observability-with-buzzfeed-and-datadog-a7e97b10424e">A Night of Monitoring and Observability with BuzzFeed and DataDog!</a> was originally published in <a href="https://medium.com/buzzfeed-tech">BuzzFeed Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DocBot: A Documentation Librarian For Slack]]></title>
            <link>https://medium.com/buzzfeed-tech/docbot-a-documentation-librarian-for-slack-part-1-bd91f64b0680?source=rss-2e461d5f22c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/bd91f64b0680</guid>
            <category><![CDATA[hackathons]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[documentation]]></category>
            <category><![CDATA[buzzfeed-posts]]></category>
            <category><![CDATA[slack]]></category>
            <dc:creator><![CDATA[Chloe (Rowan) C.]]></dc:creator>
            <pubDate>Tue, 16 Jan 2018 16:45:06 GMT</pubDate>
            <atom:updated>2018-07-31T21:28:15.339Z</atom:updated>
            <content:encoded><![CDATA[<p>As a new BuzzFeeder, one of the things I thought about often while onboarding was documentation. When using documentation to set up my environment and learn about our systems, I realized that the state of docs was inconsistent: some services had thorough documentation, other services had stale or incomplete docs. In addition to this, documentation often ended up in multiple places, or worse…deeply buried and impossible to find without the help of a co-worker.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*9y2Ojl1PJnQZc2WHjT-Jcw.gif" /></figure><p>A few months after I started, BuzzFeed threw our <a href="https://tech.buzzfeed.com/23-amazing-projects-from-buzzfeeds-2017-hack-week-d0f610dd0e94">Summer Hack Week 2017</a>. Hack Week at BuzzFeed is a time when we’re encouraged to stretch our muscles and our creativity to design hacks that we think will be fun or useful. To me it seemed like an excellent opportunity to attack the problem of documentation.</p><h3>Using Hack Week to start solving the problem…</h3><p>For Hack Week, I proposed a Slack bot project, with the goal of creating a documentation librarian. Creating a Slack integration allowed us to leverage the fact that everyone at BuzzFeed already uses Slack for office communication, and to get the project going quickly. (A Slack bot can be as simple as defining a single “slash command”, or you can build a whole web app and connect it to Slack via an API.)</p><p>During Hack Week, I worked with a small team (shout outs to summer intern Derek O’Brien and to Miguel Coquet and Arielle Benedick 👋🏼 ) to get a proof of concept up and running. Since BuzzFeed is primarily a Python and Go environment, we made the decision to create an app in Tornado (a Python framework) which would provide the greatest opportunity for others to work with us. We also decided that for proof of concept purposes we would back the app with a single SQL table, which we populated manually based on research done by team members.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qbfXtPvphCi2O8JscxahxQ.jpeg" /><figcaption>The architecture of the Hack Week version of Docbot.</figcaption></figure><p>Due to its status as a “hack,” the original Slack bot didn’t have a thorough design phase. Instead, the group creating the hack got together, brainstormed on features, and chose a handful we thought could be accomplished within the week. We also made decisions that would never fly in a production service, such as hard coding in URLs for a number of documents. These decisions helped us move quickly during the proof of concept phase, but we knew we’d end up having to change them later if this service ever made it to production.</p><p>By the end of Hack Week, we did have a working service. We could call up the readme files for any of the services listed in our table, as well as retrieve the point of contact for the service. We hooked up features to add a service’s documentation to the table and to mark whether a service was still in development or if it had been deprecated. However, the Slack bot was about to undergo an evolution…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/1*oJNmrrWnPjKpZHlolbhtAg.gif" /></figure><h3>Docbot grows up!</h3><p>Riding high on the wave of enthusiasm after Hack Week, I was ready to push this bot towards production, where it could be the integration we all dreamed of! The only problem was it needed a name and an owner.</p><p>I landed on the name Docbot as a self-evident amalgamation of Documentation and Bot. While I wish I had been more clever in naming it (I’m competing with names like “<a href="https://tech.buzzfeed.com/deploy-with-haste-the-story-of-rig-ca9a58b5719a">rig</a>,” “AV Slater,” and “Dashbird” here!) it was time to figure out where Docbot would find a home.</p><p>While I was searching for the appropriate people to help me work on this, I realized a natural fit would be the recently-created Better Docs Initiative, a working group I’m part of within BuzzFeed Tech focused on improving documentation. The group was formed by engineers who were invested in seeing the state of documentation improve at BuzzFeed. (In our documentation TODOs is an article about the Better Docs initiative!)</p><p>That meant a shift in the team: from the scrappy hack week team that got it going to folks in the Better Docs Initiative who had time to take it on. Fortunately, that transition was made easier by the fact that Arielle is also a Better Docs member, so I didn’t lose everyone from my original team!</p><p>With that settled, Docbot was ready to undergo a conversion to being production software.</p><h3>From “hack” to production!</h3><p>The first step of the process was to write a proposal for implementation. Since there had been only a makeshift design brainstorm, this proposal gave me an opportunity to think through the strengths and weaknesses of the hack, as well as the genuine needs of the software going forward. Writing out a design and implementation plan highlighted some of the trade-offs made during hacking, such as provisioning a SQL table for a data source.</p><p>A major change to implementation I made during the proposal stage was to change the data storage method. I wanted to find a way to create a “source of truth” for the bot that didn’t rely on human intervention. After surveying the landscape, I realized there was already a perfect source of truth available: our services themselves. Our code at BuzzFeed is organized in a mono repo (a single repository with a directory for each service). Each service in the mono repo must include a service definition with certain required field in order for it to be part of rig, so we added a new field. We then crawled these service directories in order to create a single API call.</p><p>The new proposed architecture of the bot looked like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*olRIG9TVGcT5pqGNUmvdbg.jpeg" /><figcaption>Architecture diagram for the proposed production version of Docbot.</figcaption></figure><p>As the biggest impact from this project would be the resources required to provision it, this proposal was delivered to the SRE team.</p><p>Once the proposal was approved, I was approached by another engineer, Ola (who’s on Better Docs with me), who wanted to know if she could use Docbot to work with a mentee on!</p><p><em>To keep in touch with us here and find out what’s going on at BuzzFeed Tech, be sure to follow us on </em><a href="http://twitter.com/buzzfeedexp"><em>Twitter @BuzzFeedExp</em></a><em> where a member of our Tech team takes over the handle for a week!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bd91f64b0680" width="1" height="1" alt=""><hr><p><a href="https://medium.com/buzzfeed-tech/docbot-a-documentation-librarian-for-slack-part-1-bd91f64b0680">DocBot: A Documentation Librarian For Slack</a> was originally published in <a href="https://medium.com/buzzfeed-tech">BuzzFeed Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Essential Elements of Giving Good Feedback]]></title>
            <link>https://medium.com/buzzfeed-tech/essential-elements-of-giving-good-feedback-3e2722565887?source=rss-2e461d5f22c2------2</link>
            <guid isPermaLink="false">https://medium.com/p/3e2722565887</guid>
            <category><![CDATA[performance-reviews]]></category>
            <category><![CDATA[buzzfeed-posts]]></category>
            <category><![CDATA[feedback]]></category>
            <category><![CDATA[advice]]></category>
            <category><![CDATA[buzzfeed]]></category>
            <dc:creator><![CDATA[Chloe (Rowan) C.]]></dc:creator>
            <pubDate>Mon, 17 Jul 2017 13:58:55 GMT</pubDate>
            <atom:updated>2017-07-17T13:58:55.423Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/715/1*NzDTOYH5mzulswui47Klig.jpeg" /></figure><p>My best friend is a hell of a woman: known for being as direct as stepping on Lego bricks barefooted, but also someone people routinely go to for advice and feedback. I spent years wondering why it was that so many people sought her out to ask her to polish their work or tell them what they were doing wrong. Finally, I couldn’t take it anymore. Late one night, while sitting on a porch in the cold, she told me we should talk about communication.</p><p>She gave me her three principles for giving feedback:</p><ul><li>Is what you’re about to say <em>honest</em>?</li><li>Is what you’re about to say <em>necessary</em>?</li><li>Is what you’re about to say <em>kind</em>?</li></ul><p>Currently, it’s mid-year review time, and I was having a conversation about gathering feedback with a fellow BuzzFeeder. I kept bringing up how important kindness is in the whole process, and they suggested I should write about it. When I sat down to write, I realized I couldn’t talk about kindness without talking about the other two aspects, so let’s dive in…</p><h3>Is it honest?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*bvN_XtzyG7lb6kMQWy7Snw.gif" /></figure><p>Even though we’re thinking about feedback right now because it’s mid-year review time, there are many times feedback happens at BuzzFeed. From interpersonal feedback (like reviews) to code reviews, from design meetings to team meetings, giving good feedback is a crucial part of the day to day here.</p><p>As a reviewer, honesty includes reviewing the feedback you’re providing. Taking the time to sit down and look through it to make sure you’re not projecting assumptions on the work in front of you is a valuable exercise. It can often reveal comments that use words like “obvious,” “easy,” “simple,” or even “basic.” These words should be seen as flags for your understanding of the problem space and the assumptions you’ve made in regards to the other party’s knowledge. Frequently, examining them reveals that the assumptions are flawed, and thus potentially intellectually dishonest.</p><p>Honesty can also mean telling someone you’re not the right person to offer feedback. Whether it’s because the thing being reviewed isn’t an area of expertise you can speak on, or because you’re not able to give it full attention, honesty requires that if you can’t engage in the review with care and intent you should disclose. It’s better for everyone involved if the feedback comes from the right people.</p><h3>Is it necessary?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*gSPTZMqMNkx-saUaVXOkUg.gif" /></figure><p>Answering the question, <em>Is it necessary, </em>helps keep scope creep in check. It also reduces the cognitive and emotional stresses of seeking feedback by reducing the number of things someone has to actually be concerned with. It is much easier to go into a feedback session knowing that you’re only dealing with feedback on the thing you brought to the table. Think of it as minimizing the scope creep of the feedback itself.</p><p>This can be hard because human nature is to believe that if you see something that can be fixed you <em>should</em> say something. Unfortunately, the tendency to want to mention everything you see can actually dilute the effect of your feedback!</p><p>How do you figure out what’s necessary then? The person asking for feedback will tell you. If they bring you a draft blog post and say, “Hey, can you look this over and help me out with it,” they’re probably not looking for advice on how to expand their thesis to cover a tangential project you’re also working on. If someone asks for a code review, it wouldn’t be the time to look at the code surrounding what they wrote and point out that someone else poorly named variables and they should go in and change it just because they’re in there.</p><h3>Is it kind?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*k8R84-V_Q7bhRmwORp1E3A.gif" /></figure><p><em>Kindness</em> often seems like the most difficult quality to exhibit in giving feedback.</p><p>When asked to give feedback, we all try to be helpful which is a major aspect of kindness. Whether it’s pointing out a mistake or weak area, or offering praise on something you’d like to see more of in the future, helpfulness in feedback is natural. I’ve definitely fallen into traps by being helpful to the exclusion of another major part of kindness: empathy.</p><p>Empathy in feedback starts with making sure that your feedback is balanced. That doesn’t mean that it has to be “all good” or even a majority good; rather, it means that your feedback calls forward the strengths you see as well as the weaknesses. Even if there are a lot of weaknesses and only a few strengths, highlighting both helps the person on the other end see where they can improve and where their skills are providing a good foundation for them to build on.</p><p>Empathy also means taking time to review your feedback before you send it. I first brought reviewing up in honesty, but it’s equally important for empathy. Make sure when you reread the feedback that you’re giving that you haven’t left comments that could be easily misinterpreted, or that seem confusing. If you review your feedback the day after you wrote it, and anything leaves you feeling like you have a question, that’s probably a place to revise.</p><p>At BuzzFeed we also exhibit another really strong part of empathy: the idea that feedback should strive to be blameless. If, on review, your feedback reads as though you are angry or disappointed, take that as a chance to drill in further and get to the real root cause. For example, feedback that someone “wrote bad tests” would go over poorly and wouldn’t lead to real change. On the other hand, revising that feedback and saying something like, “the tests for this code don’t cover the edge cases, which is an important part of testing,” gives the receiver something actionable and demonstrates that you care about their improvement, not just venting bad feelings.</p><h3>Tying it all together!</h3><p>So my best friend tells me all this about honesty, necessity, and kindness, and then she lays the secret of it on me — those are the same rules she uses for all communications. She just treats giving feedback as a communication exercise on hard mode. It completely reframed how I think about feedback and contextualized these guidelines as a framework for conversation.</p><p>Now, I approach feedback with these points in mind, and I ask myself if the things I’m about to say are honest, necessary, and kind, regardless of whether I’m the giver or the receiver. That seemingly small change in framing has taken feedback and reviews from a dreaded activity to something I look forward to, knowing I will come out of it smarter and better than I went into it.</p><p><em>To keep in touch with us here and find out what’s going on at BuzzFeed Tech, be sure to follow us on </em><a href="http://twitter.com/buzzfeedexp"><em>Twitter @BuzzFeedExp</em></a><em> where a member of our Tech team takes over the handle for a week!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3e2722565887" width="1" height="1" alt=""><hr><p><a href="https://medium.com/buzzfeed-tech/essential-elements-of-giving-good-feedback-3e2722565887">Essential Elements of Giving Good Feedback</a> was originally published in <a href="https://medium.com/buzzfeed-tech">BuzzFeed Tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>