3 minute read

A pretty packed issue this week, mainly due to all the excellent content from the recent Monitorama conference in Portland. As well as lots of monitoring we’ve got the usual mix of tools, workflow, culture and stories.


Devops Weekly is sponsored by Brightbox Cloud - serious UK-based cloud infrastructure from only 1.5p per hour (£10.95/month)

Start your £20 free trial now: http://brightbox.com/devopsweekly


A presentation that frames Devops as a banner for the creative disruption going on at the moment around IT. Lots of interesting stuff here, I’m hoping that I can find a video of this talk.

Cron is one of those easy to use but potentially hard to manage tools. This presentation introduces Crontabber, a python tool intended to make dependency management, testing, visualisation and just writing cron jobs easier.

Starting out with a very clear description of the what and the why of continuous delivery this presentation covers pipeline design focused on testing. Rather than talk about tools it focuses nicely on patterns and principles.

A very nice post about the importance of designing code for maintainability, specifically paying attention to what diffs of changes look like. The examples are in Chef but the theory is sound for anything that’s declarative.

A good topic of conversation when it comes to growing teams and codebases is how to manage change. Pull requests are the default GitHub answer but this post makes some good points about the advantages of tools like Phabricator for different workflows.

A presentation introducing the Metrics 2.0 standardisation effort. Covers the problem (naming things is hard) and lays out the proposed solution and some of the tools already using this

A good reminder that the complexity of combining all of the open source monitoring tools isn’t practical or likely a good idea for lots of small teams. A good example of a real world monitoring setup.

A short talk on a nifty use of Serverspec, integrating it with Sensu for monitoring. The trick using the sensu socket and creating alerts on the fly is very nice.

Lots of monitoring tools and talks are starting to take advantage of all that maths we learned at school. This talk is a nice primer, aimed at operations teams, looking at signal processing and filters.

Lots of the above posts come via the recent Monitorama conference. This next post is part observation based on the conference conversation, but also a call to arms. Are we building tooling aimed at the problems application developers have? And more specifically can we stop making everything a strip chart?

Cluster aware management tools appear to be of interest to lots of people at the moment. This post and presentation describe Skynet. It’s not open source yet but nicely describes the thinking behind building a simple tool to do this.


Target is seeking a Sr. Release Engineer to join our team. Requires BA/BS, 7-9 years of experience, and in-depth knowledge of state-of-the-art engineering technical approaches in design, build, testing, debugging problems as required by domain. Target offers a world of possibilities in an environment that will encourage you to innovate and contribute ideas. Join a company that’s always moving forward to what’s next.

Odecee are Hiring! We’re looking for Lead and Senior Build/Ops/DevOps Engineers to join our Melbourne and Sydney teams. Be an expert among experts and consult to some of Australia largest enterprises and projects. Australian citizens/residents are encourage to apply. s457 business visa sponsorship may be considered for a Lead role but only where all selection criteria are met.


DevOpsDays Brisbane, Australia is on the 25th and 26th of July at the Brisbane Convention & Exhibition Centre. Tickets are now on sale and a call for proposals is open until the 14th of June. Keynote speaker is Dr Sidney Dekker. Apparently July is the perfect time for a holiday at the close-by Great Barrier Reef!


Packetbeat looks like a very nice new open source application monitoring and packet tracing system. A lovely Kibana based interface and agents which can detect various types of traffic once installed make for a very simple getting started experience.

A neat approach to managing users in a large GitHub organisation. Provides tools for auditing users, ensuring two factor authentication is enabled and providing reporting of issues.

Managing multiple different git hooks across multiple repositories for self-hosted git repos can be a pain. cpthook aims to make that easier, moving the configuration into a simple YAML file and providing a command line to manage the hooks.