DEVOPS WEEKLY ISSUE #431 - 31st March 2019

3 minute read

I think this is the first Devops Weekly issue to feature a choose your own adventure game. If you’ve ever been on call, or have teams going on call for the first time, you should play a few rouns.

From our sponsor, VictorOps

In case you missed it, check out the recording of the recent VictorOps webinar, How to Make On-Call Suck Less. The webinar covers 5 actionable steps every DevOps team can take to improve alerting and make on-call suck less:
http://try.victorops.com/devopsweekly/webinar-make-on-call-suck-less

News

As more development teams own services in production, new folks are experiencing on-call. This game gives you a flavour of on-call incidents, tracking how long the incident takes to resolve and how much stress it causes.
http://dorothyjung.com/oncall-game/

WebAssembly has been interesting to folks in the browser for a while, but with work on WASI, a systems interface for WebAssembly, it has interesting usecases for infrastructure and systems work too.
https://hacks.mozilla.org/2019/03/standardizing-wasi-a-webassembly-system-interface/

A thought provoking post on issues caused by the split between the multiple layers of software and hardware, applying conway’s law to multi-tenant uses of the stack.
https://blog.jessfraz.com/post/thoughts-on-conways-law-and-the-software-stack/

A detailed outage report with lots of interesting information, especially for anyone running PostgreSQL at scale.
https://mailchimp.com/what-we-learned-from-the-recent-mandrill-outage/

A post on the importance of eventual consistency of the network when thinking about designing distributed systems. Tradeoffs between performance and simpler operations vs a more complex data model to reason about.
https://blog.envoyproxy.io/embracing-eventual-consistency-in-soa-networking-32a5ee5d443d

A post comparing the current situation with AWS and Kubernetes to how things played out with Windows and Linux.
https://zwischenzugs.com/2019/03/25/aws-vs-k8s-is-the-new-windows-vs-linux/

A presentation looking at the fundamentals of instrumenting your software for easier operation in production.
https://www.slideshare.net/postwait/put-some-sre-in-your-shipped-software

A discussion of the value and the cost of implementing a configuration management database, from a traditional ITSM point of view.
http://itsmtransition.com/2019/03/is-a-cmdb-required-for-effective-it/

Monitoring has generally been quite an ad hoc affair, but recent efforts have started to try and strandardise different aspects. Two of those, OpenTracing and OpenCensus, are merging together, which is surely a good thing.
https://medium.com/opentracing/merging-opentracing-and-opencensus-f0fe9c7ca6f0

A quick comparison of various service mesh technologies, looking at istio, linkerd and consul connect.
https://www.cloudops.com/2019/03/comparing-service-meshes-istio-linkerd-and-consul-connect/

A new version of the Traefik reverse proxy is coming and has some useful looking features. TCP handling, cross-provider support, a Kubernetes CRD and more.
https://blog.containo.us/back-to-traefik-2-0-2f9aa17be305

A post introducing the concept of observability, and providing some opinions about how this varies from more traditional monitoring.
https://sflanders.net/2019/03/26/an-intro-to-observability/

Events

The O’Reilly Velocity Conference returns to San Jose, June 10–13. Velocity is the best place on the planet for web ops and systems engineering professionals to get expert insight on building and maintaining cloud native systems. With 4 days of practical content on DevOps, systems engineering, performance, and more, there’s something for everyone. Passes start at $796 when you use the code DEVW20 (applies to Gold, Silver, and Bronze passes). Limited offer expires May 4. Register today!
https://oreil.ly/2Uf3u63

Jobs

Expensify - We’re on the hunt for one bright soul to join our infrastructure team to help battle expense reports that suck. We guarantee you fun, an amazing opportunity to learn and teach, and the siren’s call of distant riches. With a healthy mix of cloud and owned racks, you’d work hand-in-hand in our engineering org on a profitable, well-loved product with a focus on improving scaling and security.
https://we.are.expensify.com/jobs

Tools

Lucet is a WebAssembly compiler and runtime, intended to support using WebAssembly outside the browser environment.
https://www.fastly.com/blog/announcing-lucet-fastly-native-webassembly-compiler-runtime
https://github.com/fastly/lucet

Screwdriver is a relatively new build platform designed for continuous delivery. It features a pluggable execuror model with support for Docker, Nomad and Kubernetes.
https://screwdriver.cd/
https://github.com/screwdriver-cd/screwdriver

In case you missed it, check out the recording of the recent VictorOps webinar, How to Make On-Call Suck Less. The webinar covers 5 actionable steps every DevOps team can take to improve alerting and make on-call suck less:
http://try.victorops.com/devopsweekly/webinar-make-on-call-suck-less

Updated: