The theme of this article is, somebody knows. So often this is the case with lurking infrastructure issues, and it only becomes clear that somebody knew about the underlying risk once things blow up (or never). How can we find out these things that someone already knows, soon enough to act?

  Elizabeth Ayer

In this air crash investigation report, somebody knew: the maintenance supervisor had written multiple memos about a risky maintenance practice to no avail, and the practice directly contributed to the crash.

  Admiral Cloudberg

And in this one, somebody knew too: a trained pilot in a nearby village called air traffic control to warn them that a plane looked likely to crash into a mountain and needed to pull up — shortly before it hit the mountain.

  Admiral Cloudberg

A lolsob-worthy comment on laying off SREs. And here‘s a totally on-point reply with the somebody knew moment.

Partly, it’s about accepting that this is hard work. The other part is choosing where your energy input can yield the most learning.

Full disclosure: Fred is my teammate at work.

  Fred Hebert

Check it out, the incident.io folks started a podcast about incidents!

  incident.io

Here’s Google’s report for a BigQuery outage that occurred on October 13.

  Google

At last9, we auto-delete slack messages after 2 days on all personal Direct Messages. These retention policies force teams to improve documentation, kill tribal knowledge and drive accountability for mistakes, errors.

  Nishant Modak — Last9

There are some interesting tidbits in the pile of incidents in this report.

  Jakub Oleksy — GitHub

Don’t beat yourself up! This is like another form of blamelessness.

  Robert Ross — FireHydrant + The New Stack

In this article, I will share with you how setting up passive guardrails in and around developer workflows can reduce the frequency and severity of incidents and outages.

  Ash Patel — SREPath

This conference talk summary outlines the three main lessons Jason Cox learned as director of SRE at Disney.

  Shaaron A Alvares — InfoQ

Here’s a look at how Meta has structured its Production Engineer role, their name for SREs.

  Jason Kalich — Meta

Bit-flips caused by cosmic rays seem incredibly rare, but they become more likely as we make circuits smaller and our infrastructures larger.

  Chris Baraniuk — BBC

Cloudflare shares details about their 87-minute partial outage this past Tuesday.

  John Graham-Cumming — Cloudflare

In reaction to a major outage, these folks revamped their alerting and incident response systems. Here’s what they changed.

  Vivek Aggarwal — Razorpay

The author of this post sought to test a simple algorithm from a research paper that purported to reduce tail latency. Yay for independent verfication!

  Marc Brooker

In this story of SLOs gone bad, error budgets and code freezes provided a perverse incentive that caused a great deal of harm.

  dobbse.net

This article seeks to apply SRE principles to security in the form of a Threat Budget.

  Jason Bloomberg — Intellyx

After talking to hundreds of engineers about their processes, we’ve identified five of the most common challenges we see across companies looking to put more structure behind how they manage their incidents.

  Mike Lacsamana — FireHydrant

The Analysis section has a lot of important lessons. What really stands out in this incident review is the fact that Honeycomb plainly lays out the fact that they don’t yet know what went wrong, and why not.

  Fred Hebert — Honeycomb
  Full disclosure: Honeycomb is my employer.

several, small staging clusters—each fit for their purpose—offers a more maintainable, cheaper alternative.

  Tyler Cipriana

I’m really enjoying the Admiral Cloudberg series of aircraft accident investigation reports. How did I not know about these before??

A lot has improved in aviation safety since this crash in 1967, but there’s still a lot we can learn in SRE even now. For example: the operator’s view into the system should make the result of their inputs clear.

  Admiral Cloudberg

An unannounced (maybe inadvertent?) breaking change in an Azure API caused an outage. Here’s the story of the investigation.

  Nikko Campbell — Metrist

Another Admiral Cloudberg air accident investigation, this time showing how easily critical details can slip through the cracks.

  Admiral Cloudberg

This article explores the advantages of powering SLOs with observability data.

  Pierre Tessier — Honeycomb
  Full disclosure: Honeycomb is my employer.

As the James Webb Space Telescope moves into normal operations, there are more great SRE lessons to be learned.

  Jennifer Riggins — The New Stack

During 5 years of experience as an SRE, the author of this article gathered a set of best practice patterns for software development and operation, which they share with us.

  brandon willett

How Airbnb built a persistent, high availability and low latency key-value storage engine for accessing derived data from offline and streaming events.

  Chandramouli Rangarajan, Shouyan Guo, Yuxi Jin — Airbnb

By owning and reporting MTTR, teams have no choice but to be accountable for the reliability of the code they write. This dramatically changes the culture of engineering.

  Sidu Ponnappa — Last9

I learned about plan continuation bias while reading this air accident report, and I’m certain I’ve experienced this during incidents I’ve been involved in.

  Admiral Cloudberg

As a television broadcaster, how do I ensure that my channels are playing out the right thing for my viewers?

This is SRE applied to tv broadcasting: they replaced human monitoring of screens with an automated system.

  Jeremy Blythe — evertz.io
  Full disclosure: Honeycomb, my employer, is mentioned.

An interview with an engineer about on-call practices, training folks for on-call, and chaos engineering.

  Elena Boroda — Fiberplane

SRE: totally defined. Time for a reorg, and with a catchy tune!

  Forrest Brazeal

Great advice for incident response, backed up by real-world anecdotes.

  Audrey Simonne — DZone

There’s a lot to learn from in this air accident. A chilling example: several quirks of the plane’s automation combined to effectively tell the pilot to continue pushing the plane to stall.

  Admiral Cloudberg

When sharding a database, if transactions can span shards, then it can be very difficult to reason about the system’s maximum throughput.

For example, splitting a single-node database in half could lead to worse performance than the original system.

  Marc Brooker

Through Ubuntu’s unattended-upgrades system, a systemd update was installed that broke systemd-resolved, which in turn broke GitHub Codespaces. The systemd bug report they link to is also well worth a read.

  Jakub Oleksy — GitHub

Why not?

we’re, unfortunately, too good at explaining away failures without making any changes to our priors.

  Lorin Hochstein