SRE Weekly Issue #155

A message from our sponsor, VictorOps:

Machine learning and AI are becoming integrated into numerous services and applications across industries. See how SRE and DevOps teams can leverage MLOps to help shorten the incident lifecycle and maintain highly reliable systems:

http://try.victorops.com/sreweekly/mlops-incident-lifecycle

Articles

A developer’s perspective on why being on call is important and how to structure it fairly (hint: compensation).

Henrik Warne

The Conclusion section sums it up nicely:

In this post, we talked about various delivery guarantee semantics such as at-least-once, at-most-once, and exactly-once. We also talked about why exactly-once is important, the issues in the way of achieving exactly-once, and how Kafka supports it out-of-the-box with a simple configuration and minimal coding.

Rahul Agarwal — DZone

This is a riveting discussion about retrospective analysis of incidents, hosted by Microsoft. Throughout the discussion, there’s an emphasis on learning from incidents as opposed to simply coming up with action items.

Note: one of the panelists is my fellow employee at Fastly.

Jessica DeVita — Microsoft, with Duck Lawn (Pushpay), Tom Griffin (Pushpay), Sue Allspaw Pomeroy (Fastly), John Allspaw (Adaptive Capatacity Labs) and Dr. Richard Cook (Adaptive Capacity Labs)

If you’re looking for a blueprint of how to structure your SRE organization’s meetings, this is a great resource.

Dave Mangot

This post is the second part of the series on Designing Resilient Systems. In Part 1, we looked at use cases for implementing circuit breakers. In this second part, we will do a deep dive on retries and its use cases, followed by a technical comparison of both approaches.

This article is really thorough and includes a section on combining retries with circuit breakers.

Corey Scott — Grab

The problem is that most advice how to “get design right” only applies to design inside a process boundary. Most of those advices do not work well if applied to distributed systems.

What I have learnt over time is that we basically need to re-learn how to design systems, i.e., how to spread the functionality in a distributed environment.

Uwe Friedrichsen — InfoQ

This really stood out to me:

In practice, we have fixed whole classes of reliability problems by forcing engineers to define deadlines in their service definitions.

Ruslan Nigmatullin and Alexey Ivanov — Dropbox

Outages

Updated: January 13, 2019 — 8:25 pm
SRE WEEKLY © 2015 Frontier Theme