2 minute read

Everything from command line user interface design to platform-driven reorganisations and discussion of approaches to cloud adoption to talk of the benefits and risks of applying automation to testing. Definitely lots of good pieces to read this week.


Downtime can be expensive. Download the Incident Management Buyer’s Guide to learn what your DevOps team can do about it. Get your copy here:


A Tale of Two Pipelines: To DevOps or Not to DevOps

Need to increase innovation and the speed of your software delivery pipeline? Brett Hofer, Global DevOps Practice Lead at Dynatrace, will show you how through a tour of two pipelines. You’ll get insights on key pipeline factors that plague traditional delivery pipelines and how to overcome them with practices that establish a true DevOps culture for a faster and less troublesome continuous delivery pipeline.


A good discussion of building tools to enforce standards and patterns for managing multi-tenant cloud environments. The discussion of the Chandika service registry is particularly interesting.

A good presentation on the evolution of operations at Spotify. Includes a good list of expected operations activities.

A good argument for the need for more application layer security controls as many organisations abstract themselves further away from the underlying infrastructure.

Demonstrating hello-world examples for tools aimed at complex distributed systems can be tricky. Ideally you’d have a realistic demo example composed of several services - enter sock shop.

With so many of the tools we use being command line based, any discussion of improving the user experience of CLI tools is a good thing in my book. Some interesting ideas here about a formal grammar and about onboarding new users.

A look at the medium/long term impact of containers and orchestrators on typical datacenter operations - and the disruptive potential of consistent APIs for heterogeneous workloads.

How adopting a platform changes individual roles within an organisation is a common devops topic. This presentation is a detailed and very visual look at one approach, focusing on a platform and separate application team model.

An interesting approach to profiling any arbitrary code, using the new sysdig tracer functionality. If you could print to /dev/null you get a powerful tool for profiling and visualising traces.

Moving from statically addressable systems to scheduled containers poses some questions about service discovery. Here’s a simple discussion of the problem and a solution for erlang processes on Kubernetes.

A good example of using Prometheus to monitor containers, in this case those making up Kubernetes. Examples of queries and of the resulting graphs and dashboards.

A discussion of the role of automation in testing, and in particular why building tools to augment and ease testing by people, rather than to replace it, might be the right approach.


Pshtt is a handy tool for anyone managing a large number of sites and trying to enforce HTTPS best practices like use of HSTS or specific cypher suites.

Downtime can be expensive. Download the Incident Management Buyer’s Guide to learn what your DevOps team can do about it. Get your copy here: