2 minute read

Several posts this week on automation and engineering efforts with physical hardware, as well as posts on serverless, unikernels and containers. Generalising when it comes to technology in this space is dangerous, which makes it all the more impressive that the culture and organisation aspects of devops appear to be so generally applicable.


[WEBINAR] The Do’s and Dont’s of Post-Incident Analysis. Join VictorOps and Datadog to get an inside look at how to conduct modern post-incident analysis.

Sign up now: http://bit.ly/2d0Ejyk


Top Lessons Learned While Researching and Writing The DevOps Handbook

Join The DevOps Handbook co-author Gene Kim as he shares first-hand insights and lessons learned during the five years required to create this essential resource, including DevOps transformation case studies around continuous integration and delivery. Register today!


Windows 10 is an increasingly interesting development environment. One of the things I’ve been waiting for is the ability to run windows hyper-v containers and linux containers side by side. Here’s how to do it.

A good reminder that adopting devops means acting to improve the culture of your organisation, and some good tips for building bridges between former silos.

Most talk of unikernels has been about theory, but these notes from a recent talk discuss getting them into production. Lots of talk of scaling, hosting, testing and general observations of an early adopter.

An argument that end-to-end testing is mainly not worth the effort due to long lead times and maintenance costs, and a more balanced continuous testing approach is more useful.

As seen in the post above, the best way to test web application is still a matter of opinion. This survey is collecting some data surrounding application testing and I’d be interested in seeing the results.

Some good practical tips for anyone getting started with writing incident postmortems focused on learning rather than blame.

Physical infrastructure doesn’t have to mean a lack automation. A solid post on mosting to a fully automated provisioning system using CIMC and Foreman.

I attended the recent Operability conference in London and these two posts nicely summarise the various talks and provide lots of links to the slides.

An interesting post about rebuilding a common part of a web application stack at scale, in this case how GitHub rebuild their load balancer layer. Good explanation of the design goals.

In 1.12 Docker added the HEALTHCHECK instruction to the Dockerfile syntax. This post covers what it does and why, and shows how you can add healthchecks to your images.

A detailed slide deck looking at building better security monitoring into your applications. Looks at data collection, visualisation and analysis of the collected data.


LambCI is a CI and build system designed to run on AWS Lambda. This makes for some interesting properties, including potentially large cost savings depending on your build profile.

Static site generators are nothing new, but Trestus is interesting as it’s very domain specific. It aims to build status pages, for communicating with users around outages, with the data coming from Trello. If you’re using Trello internally to manage an outage that could be a nice workflow.

SQLint is a simple SQL syntax checker, ideal for sanity checking SQL files as part of a continuous integration workflow.

[WEBINAR] The Do’s and Dont’s of Post-Incident Analysis. Join VictorOps and Datadog to get an inside look at how to conduct modern post-incident analysis.

Sign up now: http://bit.ly/2d0Ejyk