Последние посты
CatOps
21 сент., 11:15
A good friend of the channel raises funds for his new unit (a small air defense unit). They are raising funds for tools and materials for their positions.You can donate to this Monobank jar:https://send.monobank.ua/jar/4SB2PtGPzV#donations #UkraineCatOps
18 сент., 10:43
There will be no digest today because I haven't posted much this week.So, instead here is an article from Anthropic on how to deal with the strain that AI development and reviews put on CI. This is a real problem in the AI-driven SDLC, or so-called "software factories", but unlike code reviews and harnesses, it's not been talked about so much, unfortunately.#cicd #aiClaudeAgentic coding is straining CI. Here’s how we scaled test impact analysis at Anthropic | Claude by AnthropicOur CI job volume increased 25x over 6 months. We patched our test selection service three times before finding a sustainable solution.CatOps
16 сент., 10:36
Why Distributed Databases Fail at Coordination Boundaries is a neat article on the areas to pay attention to when working with any distributed system, really.#dbDZoneWhy Distributed Databases Fail at Coordination BoundariesFailures in distributed systems emerge at interfaces where independent components exchange timing, ownership, and state information.CatOps
12 сент., 10:58
I don’t remember, if I shared this book bundle with you, but it won’t hurt to share it again.Think like a programmer by No Starch Press.#booksHumble BundleHumble Tech Book Bundle: Think Like a Programmer 2026 by No Starch PressThink like a programmer and move like a programmer with a collection of learning courses from the experts at No Starch Press!CatOps
11 сент., 09:41
I recall, back in the day, there was a popular interview question: "What would you not automate in your job?" I haven't heard it for a while, but with AI now, maybe it's still relevant. Perhaps it's even more relevant than back then.What SREs Should Automate — and Never Automate — with AI is an article about that. I would argue that the point about security incidents contradicts a bit to what was said earlier: if AI can collect the data for a normal incident, why cannot it do the same for the security one? A human would still stay in charge of the remediation and the final decision.#sre #aiHackernoonWhat SREs Should Automate — and Never Automate — with AI | HackerNoonAutomate based on impact and recoverability, not on whether the AI is technically capable of doing the task.CatOps
9 сент., 09:07
An interesting article on how to configure AI reviews for Terraform plans with GitHub Actions and AWS Bedrock.I don’t necessarily like the style of this article, but the content is cool. Now, I need to check how to do something similar with Atlantis. I presume, I could just trigger a GHA on a comment.P.S. This article was brought to you by weekly.tf. If you actively work with Terraform, you may find this newsletter helpful. So, make sure to subscribe!#terraform #aiMediumI Built an AI Terraform Plan Reviewer and Hit Five Access Gates Getting ThereEvery Terraform PR I review goes the same way. I open the plan output, scroll past forty lines of ~tags and ~id noise, and try to spot the…CatOps
7 сент., 08:55
Today, I’d like to remind you about the fundraiser a friend of mine is doing. The goal is to raise 70k UAH for a pickup truck for Zaporizhzhia direction.You can donate on this Monobank jar:https://send.monobank.ua/jar/8Mjotj4Uei#donations #UkraineCatOps
4 сент., 15:16
A new issue of the CatOps Digest is here!https://newsletter.catops.dev/p/catops-digest-2026-09-04#digest #newsletternewsletter.catops.devCatOps Digest 2026-09-04What was on CatOps in the last two weeks...CatOps
3 сент., 15:36
A nice visualization of Go dependencies.It's interesting how you can instantly see packages that support hundreds, if not thousands, of others. I would have guessed that testify is popular, but I wouldn't have guessed that it's the most popular.Also, I did not expect to see Viper and Cobra being so influential, even though I think they are cool.#go #programmingLuís SousaThe Shape of Go — Luís SousaEvery public Go module and every declared dependency in a single graph: 2,638,112 modules and 9,443,537 edges.1,260Открыть в TelegramCatOps
2 сент., 09:22
Continuing with the topic of AI.A colleague of mine wrote an article on how to attribute the cost with AWS Bedrock with code samples of how to enable it and the description of pitfalls we encountered along the way.#aws #ai #finopsMediumPer-engineer cost attribution for Claude Code on Amazon Bedrock: a complete guideThis mechanism applies to any Bedrock client on IAM Identity Center; our examples use Claude CodeCatOps
1 сент., 09:39
I already shared postmortems from both OpenAI and Hugging Face on the recent security incident when a frontier model broke out of its sandbox to search for a solution for its task.Here is a more detailed writeup by OpenAI on this incident. Given that OpenAI has an intrinsic motivation to advertise their models, I would still take this with a pinch of salt, but it should be interesting.I haven't read this article fully yet, but the start is promising.#ai #securityOpenAIThe Hugging Face incident and the road aheadOpenAI shares findings from the Hugging Face security incident and the steps we’re taking to strengthen AI model security, monitoring, and alignment.1,080Открыть в TelegramCatOps
31 авг., 15:04
For today's Donations Monday, I'd like to remind you about a fundraiser by a friend of mine. Its goal is to get 70k UAH for a pick up truck. So far, about 20k has been raised. Let's make it happen!https://send.monobank.ua/jar/8Mjotj4Uei#donations #UkraineCatOps
28 авг., 08:36
I've written an article about load testing of Apollo Router - a GraphQL API gateway.It's mostly pictures, so it should be easy to follow :)#graphql #load_testMediumLoad Testing Apollo RouterThis article is written for humans by humans. No AI was used in the process (yet my spelling checkers might use AI under the hood).CatOps
27 авг., 10:22
Cloudflare shares a story of how they automated GitHub issues triage with AI for their Astro project.It neatly shows the evolution from a collection of skills to a semi(?) autonomous state machine that has taken the project to zero open issues.They also share the workflows they use publicly, but in my opinion the more valuable lesson here is the step-by-step model of how they got there.#ai #open_sourceCloudflare BlogHow we built a software factory to drive Astro’s GitHub issue count to zeroBy replacing manual issue verification with isolated AI subagents running in GitHub Actions, the Astro maintainers reduced open issue count by 85%. This post explores the architecture behind automated bug reproduction, patch verification, and preview releases.CatOps
25 авг., 12:25
A great article about Postgres' MVCC, or rather, issues with MVCC and the ways other engines tried to mitigate those.It has a neat comparison of the tradeoffs, strong and weak parts of different implementations. It also provides insights into how different DB engines work.A quote:... if someone tells you a database has solved multi-versioning, ask them the four questions: where do old versions live, which way do chains point, what do indexes point at, and who cleans up. Then ask what happens when a transaction stays open over lunch.#postgres #databasesboringSQL | Supercharge your SQL & PostgreSQL powersPostgreSQL's MVCC is bad. So is everyone else's.The critics are right: PostgreSQL's MVCC bloats tables and amplifies writes. Oracle, InnoDB, SQL Server, MongoDB, LSM stores, and etcd answer the same four design questions differently, and each pays for it somewhere else.CatOps
24 авг., 15:38
Happy Independence Day of Ukraine! 💛💙It always starts with small steps, but every step counts. So today, I'd like to get back to small fundraisers that also matter!Here, a friend of mine is raising funds for a pickup track for the Zaporizhzhia direction.You can find more info in this Instagram post (in Ukrainian).You can donate to this Monobank jar: https://send.monobank.ua/jar/8Mjotj4Uei#donations #UkraineCatOps
22 авг., 11:35
A new issue of the CatOps Digest is here!https://newsletter.catops.dev/p/catops-digest-2026-08-22#newsletter #digestnewsletter.catops.devCatOps Digest 2026-08-22The last summer digest of this year...CatOps
20 авг., 13:42
University of Colorado published an article about a study of 7700 employees that shows that Remote workers report the highest well-being.The study simply analyzed the self-reported data without any judgments on why exactly this is the case, but one of the authors points out that:The study did not explore why remote workers reported higher well-being, but Johnson points to a growing body of research on autonomy and flexibility.One explanation is that remote workers have greater control over their work setup and daily schedule, she said."If you have control over your environment, you tend to have more positive outcomes," she said.Working from home can also eliminate many everyday stressors."Spending a lot of time in traffic is negatively related to well-being," Johnson said. "There are so many little stressors associated with being in the office."Those stressors can includeCatOps
19 авг., 08:59
A postmortem for the recent GitHub incident. Regarding the cause:The immediate cause of the failure was network saturation on load balancers in Central US due to a new peak in traffic. Originally this was caused by an Istio sidecar pod reaching its concurrency limits and failing to auto scale correctly because of a misconfigured policy that watched host service but not sidecar limits. One failure cascaded to more and eventually four HAProxy nodes exhausted their flow limits, degrading the gateway auth path and causing widespread authentication latency and failures. The problem was worsened by optimistic retry logic which overloaded internal load balancers. Pausing HAProxy on those nodes simultaneously produced immediate broad recovery.#postmortem #githubGithubstatusIncident with GitHub.comGitHub's Status Page - Incident with GitHub.com.CatOps
18 авг., 12:24
The Hidden Cost of Postgres Constraints at Scale is a description of a bottleneck foreign keys can create in a write-heavy PostgreSQL table. It also has recommendations of what you can do to overcome that.Now, I would say, it's a corner case for many people. The author works in a company that develops their own time series solution based on Postgres, so for them, it makes sense.For the vast majority of people, I would argue, it makes more sense to offload the integrity checks to the database, rather than trying to implement those in-house. Of course, it may be different for you, but no one knows your system better than you do.#db #postgresHackernoonThe Hidden Cost of Postgres Constraints at ScaleLearn why PostgreSQL FOREIGN KEY and UNIQUE constraints become bottlenecks at high ingest rates—and four proven ways to reduce write latency safely.
