DEVOPS WEEKLY ISSUE #246 - 20th September 2015

3 minute read

I mentioned last week it was conference season. This week’s issue features lots of content from event last week, in particular the excellent looking Automacon. Next week I’m hoping WinOps and Operability both in London have similar high quality content.

Sponsor

Devops Weekly is sponsored by Brightbox Cloud: the multi-zone cloud platform built for High Availability.

Try an SSD cloud server in 30 seconds with your £20 free credit…
http://brightbox.com/devopsweekly

Sponsored

[Webinar] Managing Technical Debt While Innovating on DevOps

How do you “keep the lights on” for your legacy systems, while also enabling your DevOps teams to launch new features faster? Join this webinar to learn tips to manage technical debt in legacy code, plus efficient ways to close loops while enabling DevOps to optimize innovation and releases. Register now:
http://ow.ly/S5xAF

News

Lots of people work in large traditional IT environments where change is hard and costly. This talk covers a number of tips for getting started with automation in those brownfield environments.
https://speakerdeck.com/stahnma/automating-in-a-brownfield-environment

As we use more and more tools to manage system our expectations of having good tools right increases. This presentation has tips; from research to design to marketing, about how to build good tools.
http://www.slideshare.net/jamtur01/developing-good-operations-tools

Some practical tips and tricks for making small Docker images even with language environments with lots of dependencies like Ruby.
http://blog.librato.com/posts/docker-images

Following up a conversation I had at the recent GOTO conference in London, this post describes some of the security characteristics of the new unikernel systems that make them interesting for anyone interested in secure system architectures.
http://www.morethanseven.net/2015/09/20/security-implications-of-unikernels/

A nice detailed post about how Twitter build a high performance replicated log service. Detailed breakdown of the problem domain and a good primer for the language around distributed logging.
https://blog.twitter.com/2015/building-distributedlog-twitter-s-high-performance-replicated-log-service

A post looking at why the current monolithic model for the Linux kernel isn’t the best one from the point of view of the user, and why microkernels or exokernels might be preferable.
http://www.wilkie.io/posts/kernels_without_abstractions_general

Another of the presentations from Automacon, entitled infrastructure as code might be literally impossible. Some good points about the inherent complexity of the things we’re managing.
http://blog.packagecloud.io/eng/2015/09/15/automacon-infrastructure-as-code-might-be-literally-impossible/

An interesting talk on the design of Sensu, in particular focusing on making configuration friendly to automation. Some good tips for anyone building modern infrastructure systems.
https://speakerdeck.com/portertech/infrastructure-as-code-and-monitoring

The Devopsdays series of events have produced a huge range of excellent content over the last 5 years, and some of the many organisers are starting to publish the audio from recorded talks in podcast form.
https://itunes.apple.com/us/podcast/devops-days-podcast/id1035163116

Speaking of Devopsdays, the recent Washington DC event had some excellent talks on everything from Lean, web performance, code reviews, security engineering to the obligatory aviation safety talk. All the videos are online now.
http://www.devopsdays.org/events/2015-washington-dc/program/

Another good talk on the evolution of a infrastructure towards the use of service discovery, in this case with Consul.
http://www.slideshare.net/dataloop/vik-bhatti-beamly-service-discovery-for-devops

Another set of excellent monitoring deep-dive posts from Datadog, this time covering how to monitor Amazon’s DynamoDB. Details of what metrics are available and how best to use them to find problems.
https://www.datadoghq.com/blog/top-dynamodb-performance-metrics/
https://www.datadoghq.com/blog/how-to-collect-dynamodb-metrics/
https://www.datadoghq.com/blog/how-medium-monitors-dynamodb-performance/

The latest version of Puppet, Puppet 4, introduced a full type system for properties. This post covers why that’s useful and shows a few simple examples using Booleans, Arrays, Integers and user defined types.
http://tuxmea.blogspot.co.uk/2015/09/puppet-4-examples-data-types.html

Devops isn’t just about tools, or only about large organisations and complex software. This post introduces some of the practices from the point of view of the Wordpress administrator.
http://erik.torgesta.com/2015/09/introduction-to-wordpress-and-devops/

Jobs

Device42 CMDB is a single source of truth for all IT Infrastructure related configuration items with simple to use REST APIs and is looking to hire its first DevOps Evangelist who will take the reins and help explain the value of our software to the DevOps community.
http://www.device42.com/jobs/devops-evangelist/?utm_source=devops_weekly

Events

The excellent All Your Base conference, all about data and databases and aimed at developers and operators, is back in London on November 13th. The schedule already looks great and the organisers are offering 5 free tickets as part of a raffle to readers of Devops Weekly.
https://goo.gl/HsDc0A
http://allyourbaseconf.com/2015/

Tools

System XVI (or S16 for short) is a new modular init system which is currently under heavy development. Might be one to watch.
https://github.com/ServiceManager/ServiceManager

Devops Weekly is sponsored by Brightbox Cloud: the multi-zone cloud platform built for High Availability.

Try an SSD cloud server in 30 seconds with your £20 free credit…
http://brightbox.com/devopsweekly

Updated: