Some solid advanced topics this week, with in-depth articles on packaging, code repository organisation and observability, and a discussion of immutable infrastructure featuring yours truly.

Sponsor

Being on-call sucks - but is it getting better? See what 800+ professionals have to say about being on-call in VictorOps’ annual “State of On-Call” report.
http://try.victorops.com/StateofOnCall/DevOpsWeekly

Sponsored

AI-Powered DevOps: Injecting Speed & Quality Across Verizon’s Cloud Pipelines

With customer experience being a top priority, Verizon turned to an automated delivery pipeline approach that leveraged cloud and container technology. This enabled teams to deploy new features faster, directly into production. But it also increased technical complexity. Learn how Verizon used Artificial Intelligence (AI)-powered monitoring, and was able to transcend this problem, and today easily manages complex, web-scale DevOps in the cloud.
http://ow.ly/I1k330elsWO

News

An excellent article on the differences of opinion, and their underlying causes, around how software should be packaged.
https://www.bassi.io/articles/2017/08/10/dev-v-ops/

Immutable infrastructure is, in my view, a great idea with a bunch of little-discussed practical considerations. I had a chat with Rob from RackN this week on the topic which is hopefully of interest to folks.
https://www.rackn.com/2017/08/09/immutable-deployment-challenges-devops/

A decent article covering using Terraform to manage a AWS Lambda function. Covers monitoring, environments, bundling dependencies and more.
https://medium.com/build-acl/aws-lambda-deployment-with-terraform-24d36cc86533

A good discussion of the practicality and costs of exactly-once message queues, summarising various other discussions and providing a good summary and observations.
http://bravenewgeek.com/fifo-exactly-once-and-other-costs/

Lots of discussion at the moment about more intelligent use of monitoring, in particular combining different sources of data to provide context to events. This post has a good example why this is so powerful.
https://blog.turbinelabs.io/beyond-observing-behavior-a0887fe0d3ca

Different ways of organising code have different tradeoffs, and warrant differentiating tooling, especially at different scales. This post is an interesting look at how the Linux kernel source code is maintained, and the people, processes and tools which contribute.
http://blog.ffwll.ch/2017/08/github-why-cant-host-the-kernel.html

A good presentation on secrets management, featuring a look at features of a good secrets solution, some of the pros and cons of using Vault, and then looking at the AWS Parameter Store.
https://docs.google.com/presentation/d/1ipP2eB9pW5j3WDvzCGz9Wy4MBoK2SMvOTZtPdf5kSNs/edit#slide=id.g1fa55b980e_0_0

An interesting post about the utility of unikernels for more than just replacing existing workloads or architectures.
http://unikernel.org/blog/2017/the-search-for-the-killer-app-of-unikernels

CNCF - Cloud Native Computing Foundation

Free Webinar - Intro to CNI, the Container Network Interface August 24, Online
http://goo.gl/dVfE6u

KubeCon + CloudNativeCon - Join leading Kubernetes, Docker, and Cloud Native technologists in Austin for a broad range of technical sessions on the cloud native ecosystem. We sold out in Berlin and are excited to see thousands of you from the community join us, this time in Austin!
http://goo.gl/wbdqve

Tools

Functions as a Service (or FaaS) is a new serverless framework aimed at running on top of Docker or Kubernetes.
https://blog.alexellis.io/introducing-functions-as-a-service/
https://github.com/alexellis/faas

OpenEBS provides containerized block storage with per-container (or pod) QoS SLAs, tiering and replica policies across AZs and environments, and predictable and scalable performance.
https://github.com/openebs/openebs

Being on-call sucks - but is it getting better? See what 800+ professionals have to say about being on-call in VictorOps’ annual “State of On-Call” report.
http://try.victorops.com/StateofOnCall/DevOpsWeekly