Tools, practices and operability articles this week. Next weekend is the annual open source extravaganza that is FOSDEM so looking forward to meeting a few folks in Brussels. If you look closely you might find me in the cafeteria writing the next issue of the newsletter.

Sponsor

Reactive, tactical, integrated, or holistic—where does your incident management fall? Read about incident management maturity to find out.
http://try.victorops.com/DevOpsWeekly/IncidentManagementMaturity

News

A well argued opinion on how organisations can actually scale transformation efforts, making the case that methodology alone is neither necessary nor sufficient. All of this is relevant to anyone looking to devops practices in large organisations.
https://dannorth.net/2018/01/26/in-praise-of-swarming/

A good post first on why you might choose a serverless platform (in this case for a machine learning problem) and then on how to go about architecting such an application.
https://tech.scrunch.com/blog/deep-learning-on-lambda-part-1-architecture-and-performance/

RabbitMQ is commonly used as a core integration technology, which makes monitoring it particularly critical. This series of posts explain what needs monitoring and how to do it.
https://www.datadoghq.com/blog/rabbitmq-monitoring/
https://www.datadoghq.com/blog/rabbitmq-monitoring-tools/

This post looks at a few common problems with application logging and how a message based approach can introduce much needed flexibility.
https://blog.rapid7.com/2018/01/16/taking-a-message-based-approach-to-logging/

A look at operability from the point of view of mechanical sympathy and performance, with a useful list of pitfalls to avoid when building operable systems.
https://medium.com/@autoletics/is-operational-sympathy-a-good-thing-ad6ae2d5f905

Even with the growth of distributions and services, people are still interested in setting up Kubernetes from scratch, I presume to help build an understanding of the various moving parts. This post is up-to-date and focused on just the needed steps.
https://blog.alexellis.io/your-instant-kubernetes-cluster/

There is quite a lot of conjecture in this post, but I agree with many of the conclusions, and it’s an interesting topic to delve into in order to understand technology adoption in general and Kubernetes in particular.
https://softwareengineeringdaily.com/2018/01/13/the-gravity-of-kubernetes/

Jobs

SNOMED International, a non profit org within the eHealth ecosystem, is hiring for a remote DevOps Engineer, based within two hours’ travel or flight time of London, i.e. UK, Ireland or EU. Joining a small DevOps team, you’ll be running live sites, ops and engineering automation, plus supporting the tools provided to the global SNOMED CT community. It’s Ansible, Jenkins, AWS, Git, Linux, MySQL/PostgreSQL. For more info, and to apply please visit:
http://techfolk.co.uk/current-jobs/devops-engineer-work-home-tl240

Tools

Akkeris is a set of components for building a platform as a service atop Kubernetes. It’s currently lacking documentation about how to set things up but might be interesting to some.
https://github.com/akkeris

Kured solves a useful problem for operators running Kubernetes clusters, namely ensuring nodes are rebooted when needed. It ensures only one rebooting node at once, and will drain connections before rebooting.
https://github.com/weaveworks/kured

EC2 Parameter Store is a useful service for anyone using AWS, and pssh is a handy tool for interacting with it. It provides a file-system-like interface, allowing for moving around and interacting with the parameters in the parameter store.
https://github.com/kountable/pssh

Reactive, tactical, integrated, or holistic—where does your incident management fall? Read about incident management maturity to find out.
http://try.victorops.com/DevOpsWeekly/IncidentManagementMaturity