DEVOPS WEEKLY ISSUE #497 - 5th July 2020
A wide mix of posts this week, from academic papers, reference architectures for Serverless, embracing policy controls in the development process, chaos engineering and lots more.
env0 sponsors Devops Weekly
Are you running Terraform locally? Want an easy way to automate it in a collaborative way? env0 is the easiest way to get started automating Terraform, giving you consistent templates, total visibility of what’s running, and automatic project level cost visibility.
Check it out at https://www.env0.com/automate-terraform
Dashboards aren’t just used in computer operations, and we can always learn from other disciplines. This excellent essay on the history of urban dashboards is well worth reading.
A project to describe a set of reference architectures for AWS Serverless applications. Lots of hard won knowledge here. Starting with a simple web service, scalable webhook and strangler pattern.
An interesting paper on Residuality Theory, and the design of complex software systems. In particular looking at non-functional properties as first class citizens of design efforts.
A nice example of a team using Open Policy Agent and Conftest to enforce best practices and various policies when authoring Kubernetes configuration.
Notes from the recent HashiConf Digital. Lots of news and case studies around Terraform, Consul, Vault and more.
A nice introduction to chaos engineering, putting it in context with other types of testing.
A presentation on the state of serverless, with lots of thoughts about where serverless is heading and what the most interesting problems are in 2020.
A look at the Azure Service Operator, a set of custom resource definitions for Kubernetes for managing Azure services like Storage Account, Virtual Machines and Azure SQL,
Vector is a lightweight, ultra-fast tool for building observability pipelines. Take data from files, syslog, statsd and more, then transform and output to S3, elasticsearch, prometheus, etc.
Ortelius is a tool for mapping microservices. Version and track application configuration and help with adopting domain driven design.
BeeMesh is intended to run services on a cluster of machines, using a peer-to-peer model and podman under the hood. Potentially interesting in edge scenarios.