2 minute read

Lots of good posts this week but a couple in particular stood out; on writing better alerts and creating more useful log files. Even with all the new technology there is still so much to learn and improve around the fundamentals.


2016/17 State of On-Call Webinar (with DevOps.com): Register to learn what 800+ survey respondents have to say about life on-call.


Gene Kim shares top lessons learned co-authoring The DevOps Handbook [Webcast Replay]

In this on-demand webcast, DevOps researcher & author, Gene Kim, shares first-hand insights and lessons he learned while co-authoring The DevOps Handbook, including: DevOps transformation case studies around continuous integration and delivery, and concrete techniques to build a culture of continuous experimentation and learning – including those from Google, Etsy, Nordstrom, and Capital One. Check it out!


Most monitoring systems will issue alerts, but how well written or designed are those alerts really? This post delves into how to write good security alerts, discussing alert levels, automatic remediation, runbooks tied to alters and more.

A presentation that sets out to explore the role of Platform as a Service in a container world. A strong argument for many parallel domain-specific PaaS’s in the future.

Presentation from the recent Config Management Camp in Berlin, all about how Cern run there large clusters. Lots of Puppet details, but also a look at tooling, monitoring, developer workflow and more.

An interesting post on the changes in operations as organisations adopt more technology, especially around changes to traditional operations teams and what that means for software engineering teams that work alongside them.

Logs are one of those things everyone has, but making them truly useful can be a challenge. This post talks about an approach called canonical log lines which aims to bridge the gap between pure metrics and raw log traces.

Really interesting notes from a lone tester at a devops conference. I’d definitely encourage people to go to at least one event outside their strict role definition if you can do - whether that’s programming, security, ITSM, testing, etc.

An interesting look at what devops means in the context of HPC. Some interesting differences as well as similarities. The point about different tooling and terminology potentially leading to silos is well made.

I’m a big fan of corporate archeology, understanding exactly why processes are in place in order to work out whether they are still addressing the original need. This post is a good example.

An interesting history of the development and scaling of a payments platform, from small beginnings to large scale international usage. Some good technology and architecture discussions in the post.

A handy reference to adding capabilities to containers, in particular looking at all the things that have ended up in SYS_ADMIN.


runv is analogous to the runc container runtime, except runv runs your container on top of a real hypervisor. Currently supports kvm, xen and virtualbox. Useful if you need additional isolation guarantees.

Funker is an interesting experimental approach to packaging up pieces of your application (functions or otherwise) as Docker containers and have them run on-demand on a swarm.

ProxySQL is a high performance MySQL proxy which allows for query caching, routing queries to different clusters, smart query firewall rules and more.

2016/17 State of On-Call Webinar (with DevOps.com): Register to learn what 800+ survey respondents have to say about life on-call.