In this blog post, we describe the journey DoorDash took using a service mesh to realize data transfer cost savings without sacrificing service quality.

  Hochuen Wong and Levon Stepanian — DoorDash

When just a few “regulars” are called in to handle every incident, you’ve got a knowledge gap to fill in your organization.

  David Ridge — PagerDuty

Dropbox expands into new datacenters often, so they have a streamlined and detailed process for choosing datacenter vendors.

  Edward del Rio — Dropbox

This is either nine things that could derail your SRE program, or a list of things to do with “not” in front of them — either way, it’s a good list.

  Shyam Venkat

We need enough alerting in our systems that we can detect lurking anomalies, but not so much that we get alert fatigue.

  Dennis Henry

A post about the importance of product in SRE, and how to make product and SRE first-class citizens in your Software Development Lifecycle.

  Jamie Allen

A relatively minor incident took a turn for the worse after the pilots attempted a close fly-by in an attempt to resolve it. I swear I’ve been in this kind of incident before, where I took risks significantly out of proportion to the problem I was trying to solve.

  Kyra Dempsey (Admiral Cloudberg)

I’ve occasionally wondered what’s behind Slack’s /remind or “clear my away status after my vacation ends”. Now I know!

  Claire Adams

This article is an exploration of consistency and coordination in distributed systems, with lots of really interesting examples.

  Lorin Hochstein

Lots of good stuff in here, including infrastructure, monitoring, and incident management tools.

   saifeddine Rajhi

my first conference

Whew, way to dive into the deep end!

  Mike [surname unknown] — SREZone

This article explains why circuit breakers are especially useful in microservice architectures based on Lambda. It explains how to implement circuit breakers using Step Functions.

   Satrajit Basu — DZone

Definitely some interesting (and spicy!) takes in this one.

  Code Reliant

When you’re at LinkedIn’s scale, building an automated abuse mitigation means designing for high throughput. The answer: lots of caching.

  Amit Mathapati — LinkedIn

A short but thought-provoking article about where SREs belong in the management heirarchy, and why.

  Jamie Allen

This is either a set of SRE interview topics or the squares for the SRE bingo card.

  Lorin Hochstein

Blame awareness only works if you work towards blame awareness with all incidents, not just the ones that affect you.

  Will Gallego

a brief history of our pipeline and the platforms, why the rebuilding was necessary, what these new services look like, and how they are being used for Netflix businesses.

  Liwei Guo, Anush Moorthy, Li-Heng Chen, Vinicius Carvalho, Aditya Mavlankar, Agata Opalach, Adithya Prakash, Kyle Swanson, Jessica Tweneboah, Subbu Venkatrav, Lishan Zhu — Netflix

Here are five concrete tips to fix your alerts and improve alert fatigue.

  Candace Shamieh, Daljeet Sandu, and Nicolas Narbais — Datadog

This article contains guidelines for many kinds of reviews and activities SRE can do to improve reliability, such as SLO reviews, dependency reviews, and more.

  Jamie Allen

However, the reality of alerting in a socio-technical system must cater not only to the mess around the signal, but also to the longer term interpretation of alerts by people and automation acting on them. This post will expand on this messiness and why Honeycomb favors an iterative approach to setting our alerts.

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

This far-ranging conversation covers many aspects of developing a reliable platform for engineering. There’s a text summary if audio’s not your thing.

  Ash Patel — SREPath

Spurred by a single-AZ outage that took down their service, Slack set out to break their system into isolated segments so that an AZ can be drained of traffic quickly and without impacting customers.

  Cooper Bethea — Slack

If you really want to understand how complex systems fail, you need to think in terms of webs rather than chains.

  Lorin Hochstein

We asked members of the PagerDuty Community what they do to remove the fear of being on-call and also asked them to share a piece of advice for those starting out on the on-call rotation and here are some of their insightful tips!

  Xenda Amici

There’s some interesting advice in here that I haven’t heard before, like rerunning the incident review meeting if you don’t get enough out of it the first time. Have any of you ever done this?

  Jonathan Word

Catchpint’s annual SRE report is out, and you can download the PDF without even having to fill out a form.

  Catchpoint

The cool thing about this article is the discussions of anti-patterns to avoid, sprinkled throughout.

  Vanessa Huerta Granda — InfoQ

I cover GCP and AWS here a lot, so now it’s Azure’s turn, with this detailed guide on load balancing.

  Shivaprasad Sankesha Narayana — DZone

Read this one to learn how Cloudflare implemented a reliable logging pipeline with 1 million log lines per second.

  Colin Douch — Cloudflare

This article describes how to clearly show your value delivered to a tech company as someone who focuses on non-functional requirements such as operability, performance, or reliability.

  Amin Astaneh — Certo Modo

Doggedly preventing a recurrence of an incident may not be the best way to protect our systems — and may in fact make things worse.

  Lorin Hochstein

Should your SLO cover a rolling 30 days? 7 days? A calendar month?

  Alex Ewerlöf

Threads was built in five months and had over 100 million users in its first week.

   Laine Campbell and Chunqiang (CQ) Tang — Meta

This article is full of advice on setting up an on-call process that’s livable and less likely to burn folks out.

  incident.io

A pilot violated a major aviation principle, and it was the right move. It’s very interesting to me that pilots are trained on the principle but not on the exceptions, with the expectation that they will react well in exceptional circumstances.

  Admiral Cloudberg

Integer IDs or UUIDs as your DB primary key? I can’t count the number of incidents I’ve been involved in where integer primary keys played a part.

  Bertrand Florat