3 minute read

A few posts this week got me thinking about the value of both reading and writing critques. It can often be hard to write, especially when you have the intention of helping improve something but aren’t sure what the solution should be. I’ll definitely be on the lookout for more good posts in this area.

Sponsor

DevOps isn’t one-dimensional. Teams implement DevOps in unique ways that are helpful to their specific causes. But, investing in the core principles of DevOps will make your team more collaborative and drive continuous improvement:
http://try.victorops.com/devopsweekly/investing-in-devops

News

If you’re reading this newsletter you’ve probably written a bunch of YAML files recently. Here are a few very useful GitHub-integrated tools for avoiding mistakes, either with invalid syntax or files that don’t match the relevant Kubernetes schema.
https://urcomputeringpal.com/2018/09/09/yaml
https://github.com/urcomputeringpal/kubevalidator

A look at the performance of the various public cloud serverless platforms, looking at concurrency and overhead across different configurations, albeit for a simple fibonacci function.
https://medium.com/elbstack/the-largest-benchmark-of-serverless-providers-ac19b55750f4

Another set of research on serverless platform performance, this time looking at cold start times. This post also explores the impact of different languages and of additional dependencies on cold start performance.
https://mikhail.io/2018/08/serverless-cold-start-war/

A useful set of in-depth posts on all things containers. Everything from namespaces, cgroups to users and networking. Lots of helpful examples used to explain the inner workings.
https://pierrchen.blogspot.com/2018/08/understand-container-index.html?m=1

Systems programming is one of those terms that is often associated with some operations and infrastructure work, but you’d probably get different definitions from different people. This post digs in to that conundrum and why we maybe need different terminology.
http://willcrichton.net/notes/systems-programming/

A presentation on the benefits of functional programming when applied to configuration, and in particular looking at the dhall configuration language.
https://docs.google.com/presentation/d/1Mmz7Fe_BKHm68DFlUjrE9Lub7DLePurrhh0CIyiuECM/edit#slide=id.p

Learning how to best utilise and engage with open source and the open source community is, I think, part of many large devops transformation. This presentation tells a story of one traditional organisation starting down this path.
https://www.slideshare.net/dberkholz/open-source-open-community-at-a-100yearold-company

A nice trick for accessing private repositories or other private resources when building Docker images by using multi-stage builds to ensure you don’t leak sensitive information.
https://jessicadeen.com/tech/how-to-access-private-azure-devops-repos-from-a-dockerfile/

Observability invariably means dealing with large volumes of data, potentially with lots of producers and consumers and storage systems. This post explores the idea of an observability pipeline to manage the resulting complexity.
https://bravenewgeek.com/the-observability-pipeline/

A critique of the Helm package manager for Kubernetes. I don’t agree with all of the arguments but I always learn more by reading criticism, please share good critical posts if you come across them.
https://medium.com/virtuslab/think-twice-before-using-helm-25fbb18bc822

Events

O’Reilly Velocity Conference Sep 30-Oct 3 in New York City

Velocity is heading to New York City in just a couple of weeks and where will you be? At the office, wishing you weren’t missing the leading conference on DevOps, performance, cloud infrastructure, and more? With Pavilion Plus passes starting at only $149, you can’t afford to miss it. Plus, as a DevOps Weekly reader, you can get 20% off of your Gold, Silver, or Bronze pass with code DOW20—that’s up to $519 in savings! Find the pass that’s right for you:
https://oreil.ly/2tC4z9q

DevOpsDays Kansas City is coming up on the 17th and 18th of October. Talks on incident response, adopting common tooling, the perils of org charts and the usual open spaces. You can get a discount of $25 with the code DEVOPSWEEKLY.
http://devopsdayskc.org/

Tools

Another Kubernetes configuration tool, Mortar supports templating with ERB and an overlay feature which allows for better variant management.
https://github.com/kontena/mortar

DevOps isn’t one-dimensional. Teams implement DevOps in unique ways that are helpful to their specific causes. But, investing in the core principles of DevOps will make your team more collaborative and drive continuous improvement:
http://try.victorops.com/devopsweekly/investing-in-devops

Updated: