DEVOPS WEEKLY ISSUE #413 - 25th November 2018

2 minute read

A relatively short issue this week due to general business, but several interesting long reads and a number of new tools for building container images, managing various flavours of service mesh and testing CloudFormation templates.

From our sponsor, Victorops

ChatOps is the result of combining human collaboration and automation in chat applications. See how ChatOps tools and technology can automate tasks and bridge the gap between productive DevOps workflows and simple chat:
http://try.victorops.com/devopsweekly/chatops-leads-to-devops

News

A detailed look at what happens as organisations adopt faster rates of change, and what that means for improving stability. Coming back to a concept of safety and the importance of fundamentally understanding the complex systems we manage.
https://m.subbu.org/taming-the-rate-of-change-439e3dccbb5d

A nice walkthrough of analysing and then shrinking the size of a Python based Docker image, reducing the size by more than 3x.
https://simonwillison.net/2018/Nov/19/smaller-python-docker-images/

An interesting post introducing the concept of data rigidity. The post uses this to explain why, in some cases, relational databases can lead to coupling data and code, and why that’s a particularly bad fit for serverless applications.
https://medium.com/@PaulDJohnston/on-serverless-and-data-rigidity-9a77eb56d73a

All of the videos from the recent Devops Enterprise Summit are now available. A wide range of content, with lots of real-world case studies in large organisations.
https://www.youtube.com/playlist?list=PLvk9Yh_MWYuxL-HzDKmTYclLvs7dCUNav

A useful review of the current state of Helm, the package manager for Kubernetes. Explains the basic moving parts, but also brings opinions from using it and how the project is evolving with Helm 3.
https://rhonabwy.com/2018/11/22/review-of-using-helm-to-package-and-host-applications/

A post (from a certain viewpoint) about some of the downsides of adopting specific devops practices. I think the first point about on-call is a case of local vs global optimisation. Someone suffered these problems before but because they were not part of the team it was easier to ignore.
https://www.business2community.com/strategy/downsides-of-devops-02140033

Jobs

TAB are hiring a Lead Engineer (DevOps Specialist)! :rocket: Are you passionate about open source? And picking the best technology for the job? Are you a firm believer in infrastructure-as-code?

While we’ve promoted DevOps culture and practice for years, we’re now hiring a leadership role to take us to the next level.

Take a look here for further details of the role and how to apply:
http://bit.ly/tabdevopsweekly

Tools

SuperGloo is an opinionated abstraction layer that will simplify the installation, management, and operation of various service mesh technologies. It makes enabling functionality like encryption, telemetry, and tracing simple, minimising the overhead of managing complex configuration.
https://supergloo.solo.io/
https://github.com/solo-io/supergloo/

cfripper is a Lambda application which can analyse a CloudFormation template and check for various security-related problems. It’s intended for use in a CI/CD pipeline, and supports users writing their own checks as well.
https://github.com/Skyscanner/cfripper

Makisu is a new container build tool, designer for use in a containerized environment like Mesos or Kubernetes. It extends the Dockerfile syntax to give control over when to generate a layer in the resulting image and supports a distributed cache to speed up builds.
https://github.com/uber/makisu

ChatOps is the result of combining human collaboration and automation in chat applications. See how ChatOps tools and technology can automate tasks and bridge the gap between productive DevOps workflows and simple chat:
http://try.victorops.com/devopsweekly/chatops-leads-to-devops

Updated: