A new issue of the CatOps Digest is here!
https://newsletter.catops.dev/p/catops-digest-2026-08-22
#newsletter #digest
https://newsletter.catops.dev/p/catops-digest-2026-08-22
#newsletter #digest
newsletter.catops.dev
CatOps Digest 2026-08-22
The last summer digest of this year...
β€1
ββHappy Independence Day of Ukraine! ππ
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 #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 #Ukraine
π18β€3
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:
#postgres #databases
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 #databases
boringSQL | Supercharge your SQL & PostgreSQL powers
PostgreSQL'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.
β€3
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_source
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_source
Cloudflare Blog
How we built a software factory to drive Astroβs GitHub issue count to zero
By 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.
π1
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_test
It's mostly pictures, so it should be easy to follow :)
#graphql #load_test
Medium
Load Testing Apollo Router
This article is written for humans by humans. No AI was used in the process (yet my spelling checkers might use AI under the hood).
β€5π₯2π€£1
ββFor 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 #Ukraine
https://send.monobank.ua/jar/8Mjotj4Uei
#donations #Ukraine
β€3
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 #security
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 #security
OpenAI
The Hugging Face incident and the road ahead
OpenAI shares findings from the Hugging Face security incident and the steps weβre taking to strengthen AI model security, monitoring, and alignment.
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 #finops
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 #finops
Medium
Per-engineer cost attribution for Claude Code on Amazon Bedrock: a complete guide
This mechanism applies to any Bedrock client on IAM Identity Center; our examples use Claude Code
π3
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
Also, I did not expect to see Viper and Cobra being so influential, even though I think they are cool.
#go #programming
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 #programming
LuΓs Sousa
The Shape of Go β LuΓs Sousa
Every public Go module and every declared dependency in a single graph: 2,638,112 modules and 9,443,537 edges.
A new issue of the CatOps Digest is here!
https://newsletter.catops.dev/p/catops-digest-2026-09-04
#digest #newsletter
https://newsletter.catops.dev/p/catops-digest-2026-09-04
#digest #newsletter
newsletter.catops.dev
CatOps Digest 2026-09-04
What was on CatOps in the last two weeks...
β€1
ββToday, Iβd 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 #Ukraine
You can donate on this Monobank jar:
https://send.monobank.ua/jar/8Mjotj4Uei
#donations #Ukraine
β€2
An interesting article on how to configure AI reviews for Terraform plans with GitHub Actions and AWS Bedrock.
I donβt 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 #ai
I donβt 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 #ai
Medium
I Built an AI Terraform Plan Reviewer and Hit Five Access Gates Getting There
Every 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 theβ¦
π4β€1π1
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 β and Never Automate β 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 #ai
What SREs Should Automate β and Never Automate β 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 #ai
Hackernoon
What SREs Should Automate β and Never Automate β with AI | HackerNoon
Automate based on impact and recoverability, not on whether the AI is technically capable of doing the task.
π₯3β€1
I donβt remember, if I shared this book bundle with you, but it wonβt hurt to share it again.
Think like a programmer by No Starch Press.
#books
Think like a programmer by No Starch Press.
#books
Humble Bundle
Humble Tech Book Bundle: Think Like a Programmer 2026 by No Starch Press
Think like a programmer and move like a programmer with a collection of learning courses from the experts at No Starch Press!
π1π₯1