Several posts this week from Velocity in London, one of the best conferences I’ve been along to in a while. A great mix of technical content and talks on the social and people side of operations and software. Congrats to all the organisers this year.

From our sponsor, Victorops

DevOps teams need to organize their incident response in order to work collaboratively and build robust systems. Improve service reliability with incident commanders and centralized on-call scheduling, alert routing, and team collaboration tools. Learn more:
http://try.victorops.com/devopsweekly/incident-management-checklist

News

I spent a few days last week at Velocity in London, and saw lots of interesting talks. These two posts covers a large number of the sessions and keynotes, with detailed notes and shots of some of the slides.
https://medium.com/@AaronKalair/velocity-conf-london-2018-thursday-talks-4eef8ff4b89a
https://medium.com/@AaronKalair/velocity-conf-london-2018-friday-talks-2edc85058f22

A look at building and maintaining infrastructure by building and maintaining a large infrastructure as code codebase.
https://www.slideshare.net/brikis98/lessons-learned-from-writing-over-300000-lines-of-infrastructure-code-120597849

Another Velocity talk, this one from me, on advanced image build patterns with Docker. Looking beyond just features to ways of using multi-stage builds and other image tooling to create maintainable images.
https://speakerdeck.com/garethr/advanced-docker-image-build-patterns

gRPC is increasingly popular for inter-service communications, but a new technology comes with new error conditions. This post walks through investigating a gRPC error, and has some recommendations for anyone adopting the technology.
https://blog.hostedgraphite.com/2018/10/30/deadlines-lies-and-videotape-the-tale-of-a-grpc-bug/

Another gRPC related post, this one looking at gRPC-web, which uses the Envoy proxy to convert requests from a client (say a web browser) into HTTP/2 gRPC calls to the backend.
https://blog.envoyproxy.io/envoy-and-grpc-web-a-fresh-new-alternative-to-rest-6504ce7eb880

A detailed walkthrough of a zero-downtime deployment strategy using Terraform and Cloud Init on AWS
https://checklyhq.com/blog/2018/08/an-in-depth-look-at-100-zero-downtime-deployments-with-terraform/

A presentation (and slides) on operating Kafka at scale on top of Kubernetes.
https://speakerdeck.com/brouberol/scaling-and-operating-kafka-in-kubernetes
https://www.youtube.com/watch?v=7eEOzguzOg0

Different types of workloads often benefit from different schedulers. But at scale efficiency and resource utilisation become a problem. This post explores a custom large-scale cluster scheduler and lays out some of the tradeoffs of different approaches.
https://eng.uber.com/peloton/

A suite of tools for Linux fleet management. These include resource control, resource utilization, workload isolation, load balancing, measuring, monitoring
https://code.fb.com/open-source/linux/
https://opensource.fb.com/#linux

A rundown of various Ingress solutions for Kubernetes. Lots of options here and this post looks at 9 current options.
https://kubedex.com/nginx-ingress-vs-kong-vs-traefik-vs-haproxy-vs-voyager-vs-contour-vs-ambassador/

Jobs

Senior Devops Engineer, Remote Friendly Citrusbyte is working with Google to build a nation-wide DevOps team and has the ability to give a few individuals the opportunity to work remotely for Google. You will work on various DevOps projects internal to Google as well as deploy complex systems to GCP for Fortune 500 clients.
https://citrusbyte.applytojob.com/apply/MWcY29iFEt?source=DevopsWeekly

Tools

Garden is a new development tool for local development of multi-service applications. Re-build only the services that changed and deploy to a local Kubernetes cluster.
https://garden.io/
https://github.com/garden-io/garden

An interesting combination of using OPA (Open Policy Agent) in the context of a Kubernetes admission controller. Create policies to ensure your Kubernetes cluster meets governance or legal rules.
https://github.com/Azure/kubernetes-policy-controller

DevOps teams need to organize their incident response in order to work collaboratively and build robust systems. Improve service reliability with incident commanders and centralized on-call scheduling, alert routing, and team collaboration tools. Learn more:
http://try.victorops.com/devopsweekly/incident-management-checklist