<?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 Damien Pacaud on Medium]]></title>
        <description><![CDATA[Stories by Damien Pacaud on Medium]]></description>
        <link>https://medium.com/@damien.pacaud?source=rss-c9ccae6063d7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*-kQFve7vBySjy_fh_57sRQ.jpeg</url>
            <title>Stories by Damien Pacaud on Medium</title>
            <link>https://medium.com/@damien.pacaud?source=rss-c9ccae6063d7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 02:07:49 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@damien.pacaud/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[Scaling your on-duty team]]></title>
            <link>https://medium.com/teads-engineering/scaling-your-on-duty-team-bc467c480747?source=rss-c9ccae6063d7------2</link>
            <guid isPermaLink="false">https://medium.com/p/bc467c480747</guid>
            <category><![CDATA[scaling]]></category>
            <category><![CDATA[incident-management]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[sre]]></category>
            <category><![CDATA[team-collaboration]]></category>
            <dc:creator><![CDATA[Damien Pacaud]]></dc:creator>
            <pubDate>Tue, 16 Jan 2018 07:01:01 GMT</pubDate>
            <atom:updated>2018-02-16T07:39:49.275Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jzkj2AuRnH8jOAaM_a_TQQ.png" /></figure><p>The internet never sleeps, and even with the best design for resilience, one day, your system will go down.</p><p>At Teads, we deliver outstream video advertising for the biggest content publishers in the world. Any downtime has important repercussions on our revenue but also on the publisher’s revenue.</p><p>In a few years we grew from a start-up to a scale-up, although we operate globally, <strong>our</strong> <strong>tech team is mostly based in France</strong>. For this reason, we decided to carefully think about scaling our on-duty team in order to minimize the downtime when a system goes down.</p><p>That story is below.</p><h3><strong>On-duty in a fast-growing company</strong></h3><p>In a few years, we’ve scaled from a growing startup operating with a few pizza teams into a company where more than 100 developers on 3 different locations deliver new features on a daily basis.</p><p>We’ve been able to do so by implementing our own version of the <em>“Spotify model”</em> and it has given us the ability to stay agile while growing the tech team. Applying the same recipe to the on-duty team was a challenge, to say the least.</p><p>Initially, the on-duty team was composed of a few developers that had been with Teads since the beginning and that were very knowledgeable on every part of the platform.</p><p>We relied on their knowledge, availability and on the fact that they helped build most of the system. As we grew, the system became larger and more complex. The handful of developers keeping the revenue safe overnight were now unable to keep up with the amount of knowledge needed to solve a problem.</p><h3><strong>First step: Growing the on-duty team</strong></h3><p>We started looking for people to add to the on-duty team and ideally have someone from each of our feature teams be part of the rotation.</p><p>This was our way of implementing <em>“you built it, you run it”</em>.</p><p>It meant growing that team to 12 people and that’s when we hit the first wall.</p><p>We tried growing the team while having a few visible production incidents (<a href="https://aws.amazon.com/message/41926/">S3 Service Disruption in us-east-1</a>, anyone ?) and of course, no one was voluntarily applying to be on duty.</p><h3><strong>Lost battle: trying to be ready</strong></h3><p>One of the main reason nobody was up to the task was the complexity of our system. It can be frightening not having a complete understanding of it while having to react when an incident arises.</p><p>We tried to tackle this problem, and for a few months we set up meetings, put knowledgeable people in a room and <strong>asked them to kindly document</strong> the steps to take when incidents happen.</p><p>This was too large of a mission, even for a highly motivated team. Soon, meetings were skipped, and documentation was not improving.</p><p>At this point, we started thinking about the problem in a different way.</p><h3><strong>Enter on-duty pairing</strong></h3><p>The first decision we took was to have two persons on-duty at the same time for a week-long shift. We tried to wisely choose pairs for mutually exclusive skills set and experience. We would for example pair a back-end developer with a data-oriented developer. This allowed us to cover most systems on the critical chain.</p><p>The benefits that we see with the on-duty pairing are:</p><ul><li>It’s much easier to <strong>bounce ideas off someone </strong>when a problem is impacting production and you (or your pair) do not know how to fix it.</li><li>Sometimes while on-duty, the incident runs so deep that a critical business decision must be taken. It is <strong>much easier to share the responsibility</strong> of such a decision in the middle of the night. We accept that this may slow down the decision process as there will be back-and-forth between the pair.</li><li>In the rare event of <strong>someone not waking up</strong> to the PagerDuty calls, there is a backup. Interestingly enough, we had never experienced someone not waking up until we started pairing. This brought the question that pairing may lower each individual’s sense of alert because there is a backup, but in the end we feel it has more benefits than downsides.</li></ul><p>We implemented this change in a few weeks and so far we are quite happy with it. The team has scaled to 12 developers, coming from all feature teams, and the rotation goes smoothly.</p><h3>Escalation ?</h3><p>The traditional way of dealing with increasing complexity is to have an escalation policy. We chose not to implement this and have PagerDuty automatically wake up both pairing developers at the same time.</p><p>This automates the decision of waking-up another human being and <strong>makes PagerDuty responsible for it</strong>. We don’t want to be responsible for this hard decision so we let the robot do it.</p><p>Escalation usually also solves the <em>“I need an expert on [insert any well known distributed system here] and I need her right now”</em> problem.</p><p>Putting them on escalation policies is great if you have a big enough pool of experts on each of the systems that you use. For us, this meant that a few persons would be on call every other week. We thought this was not acceptable and decided that we could solve this by :</p><ul><li>Telling the on-duty team members we know they will do their best to recover the issue</li><li>Giving them the confidence that, <strong>as engineers, they will find a solution</strong></li><li>Automating as much as we can routine maintenance operations (taking a bad Cassandra node out of the ring, decommissioning and replacing a Kafka broker…)</li></ul><h3>Post-incident and Playbook</h3><p>Soon after the incident, we gather everyone from the on-duty team in a room for a blameless, fact-oriented, post-mortem. We aim to leave the room after one hour having filled our very simple post-incident template:</p><ul><li>Summary of the issue</li><li>How to reply to such an issue (should it rise again)</li><li>Action plan</li></ul><p>This process allows us to document our interventions and ensure, should the same incident happen again, we have a solution to mitigate its effect in a timely manner.</p><h3>Feedback</h3><p>After a few months, we are quite happy with this new on-duty rotation. It has proven useful many times and <strong>we now have</strong> <strong>more documentation than ever</strong> on how to react to our alerts.</p><p>The post-incident ritual also acts as a team bonding meeting and we are thinking of creating more rituals specifically for the on-duty team (on top of each individual’s feature team rituals).</p><p>The biggest complexity that we encountered since launching was organizing the Christmas rotation period with pairs. It’s always a challenge to find one person available during those holidays, so trying to find two is double the fun.</p><p>This article was initially published on <a href="http://sysadvent.blogspot.fr/2017/12/day-11-scaling-your-on-duty-team.html">SysAdvent</a> 2017 <em>“the annual advent calendar for SysAdmins, Ops, DevOps, and all the other folks that are excited about systems.”</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bc467c480747" width="1" height="1" alt=""><hr><p><a href="https://medium.com/teads-engineering/scaling-your-on-duty-team-bc467c480747">Scaling your on-duty team</a> was originally published in <a href="https://medium.com/teads-engineering">Teads Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Real-life AWS cost optimization strategy]]></title>
            <link>https://medium.com/teads-engineering/real-life-aws-cost-optimization-strategy-at-teads-135268b0860f?source=rss-c9ccae6063d7------2</link>
            <guid isPermaLink="false">https://medium.com/p/135268b0860f</guid>
            <category><![CDATA[cloud-computing]]></category>
            <category><![CDATA[optimization]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[infrastructure]]></category>
            <category><![CDATA[amazon-web-services]]></category>
            <dc:creator><![CDATA[Damien Pacaud]]></dc:creator>
            <pubDate>Mon, 30 Oct 2017 07:39:47 GMT</pubDate>
            <atom:updated>2018-06-04T14:40:44.497Z</atom:updated>
            <content:encoded><![CDATA[<h3>Real-life AWS infrastructure cost optimization strategy</h3><h4>How we keep a good Amazon Web Services billing hygiene at Teads</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C5SHSMbBScmcgibRTmZ-mQ.jpeg" /></figure><p><em>AWS recently </em><a href="https://aws.amazon.com/fr/blogs/aws/new-per-second-billing-for-ec2-instances-and-ebs-volumes/"><em>announced</em></a><em> its new per second billing for its EC2 instances and EBS volumes. This is perfect timing to talk about cost optimization. After a short intro we will guide you through some real world examples and best practices that we use at Teads to optimize our infrastructure costs.</em></p><h3><strong>The cloud computing opportunity and its traps</strong></h3><p>One of the advantages of cloud computing is its ability to <strong>fit the infrastructure to your needs</strong>, you only pay for what you really use. That is how most hyper growth startups have managed their incredible ascents.</p><p>Most companies migrating to the cloud embrace the “lift &amp; shift” strategy, replicating what was once on premises.</p><p>You most likely won’t save a penny with this first step.</p><p>Main reasons being:</p><ul><li>Your applications do not support elasticity yet,</li><li>Your applications rely on complex backend you need to migrate with (RabbitMQ, Cassandra, Galera clusters, etc.),</li><li>Your code relies on being executed in a known network environment and most likely uses NFS as distributed storage mechanism.</li></ul><p>Once in the cloud, you need to “cloudify” your infrastructure.</p><p>Then, and only then, will you have access to <strong>virtually infinite computing power and storage</strong>.</p><p>Watch out, <strong>this apparent freedom can lead to very serious drifts</strong>: over provisioning, under optimizing your code or even forgetting to “turn off the lights” by letting that small PoC run more than necessary using that very nice r3.8xlarge instance.</p><p>Essentially, you have just replaced your need for capacity planning by a <strong>need for cost monitoring and optimization</strong>.</p><h3><strong>The dark side of cloud computing</strong></h3><p>At Teads we were “born in the cloud” and we are very happy about it.</p><p><strong>One of our biggest pain today with our cloud providers is the complexity of their pricing.</strong></p><p>It is designed to look very simple at the first glance (usually based on simple metrics like $/GB/month or $/hour or, more recently, $/second) but as you expand and go into a multi-region infrastructure mixing lots of products, you will have a hard time tracking the ever-growing cost of your cloud infrastructure.</p><p>For example, the cost of <strong>putting a file on S3 </strong>and serving it from there<strong> includes four different lines of billing</strong>:</p><ul><li>Actual storage cost (80% of your bill)</li><li>Cost of the HTTP PUT request (2% of your bill)</li><li>Cost of the many HTTP GET requests (3% of your bill)</li><li>Cost of the data transfer (15% of your bill)</li></ul><h3>Our take on Cost Optimization</h3><ul><li><strong>Focus on structural costs - </strong>Never block short term costs increase that would speed up the business, or enable a technical migration.</li><li><strong>Everyone is responsible - </strong>Provide tooling to each team to make them autonomous on their cost optimization.</li></ul><p>The limit of cost optimization for us is when it drives more complexity in the code and less agility in the future, for a limited ROI. <br>This way of thinking also helps us to tackle cost optimisation in our day to day developments.</p><p>Overall we can extend this famous <a href="http://wiki.c2.com/?MakeItWorkMakeItRightMakeItFast">quote</a> from Kent Beck:</p><blockquote>“Make it work, make it right, make it fast” <strong><em>… and then cost efficient.</em></strong></blockquote><h3><strong>Billing Hygiene</strong></h3><p>It is of the utmost importance to <strong>keep a strict billing hygiene</strong> and know your daily spends.</p><p>In some cases, it will help you identify suspicious uptrends, like a service stuck in a loop and writing a huge volume of logs to S3 or a developer that left its test infrastructure up &amp; running during a week-end.</p><p>You need to arm yourself with a detailed monitoring of your costs and spend time looking at it <strong>every day</strong>.</p><p>You have several options to do so, starting with AWS’s own tools:</p><ul><li><a href="http://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/view-billing-dashboard.html"><strong>Billing Dashboard</strong></a>, giving a high level view of your main costs (Amazon S3, Amazon EC2, etc.) and a rarely accurate forecast, at least for us. Overall, it’s <strong>not detailed enough</strong> to be of use for serious monitoring.</li><li><a href="http://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/DetailedBillingReport.html"><strong>Detailed Billing Report</strong></a>, this feature has to be enabled in your account preferences. It sends you a daily gzipped .csv file containing one line per billable item since the beginning of the month (e.g., instance A sent X Mb of data on the Internet). <br>The detailed billing is an <strong>interesting source of data</strong> once you have added custom tags to your services so that you can group your costs by feature / application / part of your infrastructure. <br>Be aware that this file is <strong>accurate within a delay of approximately two days </strong>as it takes time for AWS to compute the files. <br><strong><em>UPDATE</em> </strong><em>(June ‘18) Detailed Billing is officially deprecated, use the </em><a href="https://docs.aws.amazon.com/awsaccountbilling/latest/aboutv2/billing-reports-costusage.html"><em>Cost and Usage Report</em></a><em> instead.</em></li><li><a href="https://aws.amazon.com/fr/premiumsupport/trustedadvisor/"><strong>Trusted Advisor</strong></a>, available at the business and enterprise support level, also includes <a href="https://aws.amazon.com/fr/premiumsupport/trustedadvisor/best-practices/">a cost</a> section with interesting optimization insights.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/784/0*x_AW2ZJNQDUL39pn.PNG" /><figcaption>Trusted Advisor cost section - Courtesy of AWS</figcaption></figure><ul><li><a href="https://aws.amazon.com/fr/aws-cost-management/aws-cost-explorer/"><strong>Cost Explorer</strong></a>, an interesting tool since its <a href="https://aws.amazon.com/fr/blogs/aws/aws-cost-explorer-update-better-filtering-grouping-report-management-ri-reports/">update</a> in august 2017. It can be used to <strong>quickly identify trends</strong> but it is still limited as you cannot build complete dashboards with it. It is <strong>mainly a reporting tool</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/0*HdSoUe7LSIP_cZuv." /><figcaption><em>Example of a Cost Explorer report — AWS documentation</em></figcaption></figure><p>Then you have several <strong>other external options</strong> to monitor the costs of your infrastructure:</p><ul><li><strong>SaaS products</strong> like <a href="https://www.cloudyn.com/">Cloudyn</a> / <a href="https://www.cloudhealthtech.com/">Cloudhealth</a>. These solutions are really well made and will tell you how to optimize your infrastructure. Their pricing model is based on a percentage of your annual AWS bill, not on the savings that the tools will help you make, which was a show stopper for us.</li><li>The <strong>open source project</strong> <a href="https://github.com/Teevity/ice">Ice</a>, initially developed by Netflix for their own use. Recently, the leadership of this project was transferred to the french startup <a href="https://www.teevity.com/">Teevity</a> who is also offering a SaaS version for a fixed fee. This could be a great option as it also handles GCP and Azure.</li></ul><h4>Building our own monitoring solution</h4><p>At Teads <strong>we decided to go DIY using the detailed billings files</strong>.</p><p>We built a small Lambda function that ingests the detailed billing file into Redshift every day. This tool helps us slice and dice our data along numerous dimensions to dive deeper into our costs. We also use it to spot suspicious usage uptrends, <strong>down to the service level</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*DCFRayCyQoC1t_60." /><figcaption><em>This is an example of our daily dashboard built with chart.io, each color corresponds to a service we tagged</em></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*p8G529QdZs9ePErk." /><figcaption><em>When zoomed on a specific service, we can quickly figure out what is expensive</em></figcaption></figure><p>On top of that, we still use a spreadsheet to <strong>integrate the reservation upfronts</strong> in order to get a complete overview and the full daily costs.</p><h3><strong>Now that we have the data, how to optimize?</strong></h3><p>Here are the <strong>5</strong> pillars of our cost optimization strategy.</p><h4><strong><em>1 - Reserved Instances (RIs)</em></strong></h4><p>First things first, you need to <a href="https://aws.amazon.com/ec2/pricing/reserved-instances/">reserve your instances</a>. Technically speaking, RIs will only make sure that you have access to the reserved resources.</p><p>At Teads our reservation strategy is based on <strong>bi-annual reservation batches</strong> and we are also evaluating higher frequencies (3 to 4 batches per year).</p><p>The right frequency should be determined by the best compromise between flexibility (handling growth, having leaner financial streams) and the ability to manage the reservations efficiently. <br>In the end, managing reservations is a time consuming task.</p><p>Reservation is mostly a financial tool, you commit to pay for resources during 1 or 3 years and get a discount over the on-demand price:</p><ul><li>You have two types of reservations, <strong>standard or convertible</strong>. Convertible lets you change the instance family but comes with a smaller discount compared to standard (avg. 75% vs 54% for a convertible). They are the best option to leverage future instance families in the long run.</li><li>Reservations come with <strong>three different payment options</strong>: Full Upfront, Partial Upfront, and No Upfront. With partial and no upfront, you pay the remaining balance monthly over the term. We prefer partial upfront since the discount rate is really close to the full upfront one (e.g. 56% vs 55% for a convertible 3-year term with partial).</li><li>Don’t forget that <strong>you can reserve a lot of things</strong> and not only Amazon EC2 instances: Amazon RDS, Amazon Elasticache, Amazon Redshift, Amazon DynamoDB, etc.</li></ul><h4>2 -<strong><em> Optimize Amazon S3</em></strong></h4><p>The second source of optimization is the <strong>object management on S3</strong>. Storage is cheap and infinite, but it is not a valid reason to keep all your data there forever. Many companies do not clean their data on S3, even though <strong>several trivial mechanisms</strong> could be used:</p><p><a href="http://docs.aws.amazon.com/AmazonS3/latest/dev/object-lifecycle-mgmt.html"><strong>The Object Lifecycle</strong></a> option enables you to set simple rules for objects in a bucket :</p><ul><li><strong>Infrequent Access Storage</strong> (IAS): for application logs, set the object storage class to Infrequent Access Storage after a few days. <br>IAS will cut the storage cost by a factor of two but comes with a higher cost for requests. <br>The main drawback of IAS is that it uses 128kb blocks to store data so if you want to store a lot of smaller objects it will end up more expensive than standard storage.</li><li><strong>Glacier</strong>: <a href="https://aws.amazon.com/fr/glacier/">Amazon Glacier</a> is a very long term archiving service, also called cold storage. <br>Here is a <a href="https://blog.cloudability.com/aws-s3-understanding-cloud-storage-costs-to-save/">nice article from Cloudability</a> if you want to dig deeper into optimizing storage costs and compare the different options.</li></ul><p>Also, don’t forget to <strong>set up a delete policy</strong> when you think you won’t need those files anymore.</p><p>Finally, enabling a <strong>VPC Endpoint</strong> for your Amazon S3 buckets will suppress the data transfer costs between Amazon S3 and your instances.</p><h4>3 -<strong><em> Leverage the Spot market</em></strong></h4><p>Spot instances enables you to use AWS’s spare computing power at a <strong>heavily discounted price</strong>. This can be very interesting depending on your workloads.</p><p>Spot instances are bought using some sort of <strong>auction model</strong>, if your bid is above the spot market rate you will get the instance and only pay the market price. However <strong>these instances can be reclaimed</strong> if the market price exceeds your bid.</p><p>At Teads, we usually bid the on-demand price to be sure that we can get the instance. We only pay the “market” rate which gives us a rebate up to 90%.</p><p>It is worth noting that:</p><ul><li>You get a <strong>2 min termination notice</strong> before your spot is reclaimed but <a href="https://aws.amazon.com/fr/blogs/aws/new-ec2-spot-instance-termination-notices/">you need to look for it</a>.</li><li>Spot Instances are easy to use for <strong>non critical batch workloads</strong> and interesting for <strong>data processing</strong>, it’s a very good match with Amazon Elastic Map Reduce.</li></ul><h4><strong><em>4 - Data transfer</em></strong></h4><p>Back in the physical world, you were used to pay for the network link between your Data Center and the Internet.</p><p>Whatever data you sent through that link was free of charge.</p><p><strong>In the cloud, data transfer can grow to become really expensive.</strong></p><p>You are charged for data transfer from your services to the Internet but also <strong>in-between AWS Availability Zones.</strong></p><p>This can quickly become an issue when using distributed systems like <strong>Kafka</strong> and <strong>Cassandra</strong> that need to be deployed in different zones to be highly available and constantly exchange over the network.</p><p>Some advice:</p><ul><li>If you have instances communicating with each other, you should try to <strong>locate them in the same AZ</strong></li><li>Use managed services like Amazon DynamoDB or Amazon RDS as their inter-AZ replication costs is built-in their pricing</li><li>If you serve more than a few hundred Terabytes per months <strong>you should discuss with your account manager</strong></li><li><strong>Use </strong><a href="https://aws.amazon.com/cloudfront/"><strong>Amazon CloudFront</strong></a> (AWS’s CDN) as much as you can when serving static files. The data transfer out rates are cheaper from CloudFront and free between CloudFront and EC2 or S3.</li></ul><h4><strong><em>5 - Unused infrastructure</em></strong></h4><p>With a growing infrastructure, you can rapidly forget to turn off unused and idle things:</p><ul><li><strong>Detached Elastic IPs</strong> (EIPs), they are free when attached to an EC2 instance but you have to pay for it if they are not.</li><li>The <strong>block stores</strong> (EBS) starting with the EC2 instances are preserved when you stop your instances. As you will rarely re-attach a root EBS volume you can delete them. Also, snapshots tend to pile up over time, you should also look into it.</li><li>A <strong>Load Balancer</strong> (ELB) with no traffic is easy to detect and obviously useless. Still, it will cost you ~20 $/month.</li><li>Instances with <strong>no network activity</strong> over the last week. In a cloud context it doesn’t make a lot of sense.</li></ul><p>Trusted Advisor can help you in detecting these unnecessary expenses.</p><h3><strong>Key takeaways</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*XA92exx-i7MgbnKhb24uIA.png" /></figure><p>Thank you for reading. This article was inspired by the talks I made during the <a href="https://www.meetup.com/Montpellier-Amazon-Web-Services/events/242281837/?eventId=242281837">#2 AWS Montpellier Meetup</a> and <a href="http://2017.devops-dday.com/">Devops D-Day</a> conference.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jHOay3TlQwv-MiFfTIlT-w.jpeg" /><figcaption>Devops D-Day 2017 — Marseille</figcaption></figure><p>If you like working on big cloud infrastructures and growth challenges, feel free to <a href="/service/mailto:mathilde.barthe@teads.tv"><strong>contact</strong></a> us, we are constantly <a href="https://teads.tv/teads-jobs/">looking</a> for great teammates.</p><p>If you want to know more about Engineering at Teads:</p><p><a href="https://medium.com/teads-engineering/about-teads-engineering-64a12cab5369">About Teads</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=135268b0860f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/teads-engineering/real-life-aws-cost-optimization-strategy-at-teads-135268b0860f">Real-life AWS cost optimization strategy</a> was originally published in <a href="https://medium.com/teads-engineering">Teads Engineering</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>