I’ve used this technique in the past with a single-page app and a highly-cacheable API, to ensure stability even when the backend goes down.

Patrick Hamann

Full disclosure: Fastly is my employer.

Here’s a deep dive into how your CA’s certificate can affect your application’s reliability — at least in the eyes of your customers.

Scott Helme

Here’s Coinbase’s followup from their outage last week.

Michael de Hoog — Coinbase

Kyle Kingsbury recently did an analysis of PostgreSQL 12.3 and found that under certain conditions it violated guarantees it makes about transactions, including violations of the serializability transaction isolation level.

I thought it would be fun to use one of his counterexamples to illustrate what serializable means.

Lorin Hochstein

Failure mode and effects analysis (FMEA) is a decades-old method for identifying all possible failures in a design, a manufacturing or assembly process, or a product or service.

If you’ve been tasked with applying FMEA in your SRE work, this article will get you started.

Matthew Helmke

This article in a nutshell:

• Nines don’t matter if users aren’t happy (h/t Charity Majors)
• Chaos engineering

Kolton Andrus — Gremlin

I hadn’t heard of this distinction before. If you haven’t either, click through to find out more.

Ayende Rahien — RavenDB

In our experience, the three big sources of production stress are:

• Toil
• Bad monitoring
• Immature incident handling procedures

Cheryl Kang — Google

ProPublica picks apart the incident in exhaustive detail, showing how multiple problems interwoven in the organization contributed to this tragedy.

Robert Faturechi, Megan Rose and T. Christian Miller — ProPublica

There’s a great review of Rasmussen’s safety boundary model, which I wasn’t previously familiar with. A system moves between three boundaries:

• the boundary to economic failure
• the boundary of unacceptable work load
• the boundary of functionally acceptable performance

Lorin Hochstein

This one includes a really nifty graph showing how reliable your N backend microservices need to be in order to hit a given reliability target R.

Bill Duncan

Here are the results of the survey I linked here a couple weeks ago. There are some interesting and surprising results, well worth a read.

Rich Burroughs — FireHydrant

A commonly-used CA’s Root certificate expired, causing some havoc. Even though Sectigo did everything right, some software didn’t handle the transition to the new root well.

Paul Ducklin — Naked Security

Casey Rosenthal tips over a herd of sacred cows with this talk that opens with 6 myths about reliable systems.

Casey Rosenthal — Verica

This is written as talking about scale during a job interview, and it’s a pretty good read even if you’re not interviewing right now.

Denise Yu

John Allspaw says we should ask “how”, not “why”. Hollnagel and Woods say that finding out why a joint cognitive system does what it does rather than how. Who’s right?

Lorin Hochstein

Yay, another issue! This one revolves around learning from incidents from organizations in other fields (Bose and NASA).

Jaime Woo and Emil Stolarsky — Incident Labs

This is a followup analysis of a Google Hangouts oiutage from last month.

Google

Catchpoint is holding a mini-conference on the ways that SRE has changed as we shift to all-remote work, and I’m super-excited to be on the Q&A panel! Hope to see you there.

Catchpoint

A seasoned pro discusses some pitfalls of cloud-based architecture based on hard-won experience.

Rachel by the bay

Monzo is back with updates on how their on-call has changed since their original article in 2018.

Shubheksha Jalan — Monzo

Along with this rockin’ article about why it’s important to make on-call bearable, Incident Labs also has a survey on your on-call experience. Click through for the link.

Incident Labs

This really crystallizes a lot of my concerns with anomaly detection.

Danyel Fisher — The New Stack / Honeycomb

If you ask someone why they did something, they’re likely to invent a logical-sounding reason without meaning to.

Lorin Hochstein

Check out this new 100-page ebook on incident response from Atlassian, great for folks setting up a brand new on-call structure or improving their existing one. It even has a section on compensating teams for being on-call.

Serhat Can — Atlassian

Laura Maguire discusses the compelling data from her PhD dissertation that the Incident Command System actually makes incident response less efficient, along with lots of other interesting findings.

Laura Maguire

A summary of a great talk by Amy Tobey at Failover Conf, amusingly framed as a “retrospective”.

Hannah Culver — Blameless

In this case, the “cloud” refers to actual clouds, the ones in the sky. It’s a comparison between concepts in aviation and SRE, fields that have significant overlaps.

Bill Duncan

My favorite:

The fact that you need to make changes to maintain availability, will itself threaten your availability.

Lee Atchison — diginomica

A bug in a new release of the Facebook SDK caused some iOS apps to crash.

Brian Barrett — WIRED

[…] I’m no longer going to try do get engineers to stop saying root cause while they are engaged in an incident. If the experts at anomaly response find it a useful phrase while they are doing their work, we should recognize this as a part of their expertise.

Loren Hochstein