Continuing with the theme of Kubernetes, I have recently built out a solution to inject environment variables into containerised applications from Hasicorp Consult and Vault Key Value (KV) engine, which might be considered as a first step in realising Hashicorp’s Service Mesh. Installing both Consul and Vault via helm with the KV Engine is fairly straightforward. Supplying these KV’s as environment variables to the containerised applications in Kubernetes, however, requires a bit more thought. Two different approaches are required to lift in values from Consul and Vault which makes things even more interesting. The approach I took was to write the KV’s to file before they are exposed as ENVs in the container, which is less than ideal. As a side note, it might be cleaner and simpler to manage config at the application layer by calling Consul and Vault’s HTTP API. That is another approach which I’m not going to talk about here.
Nov 19, 2021
If you’re managing a data engine inside a kubernetes cluster then implementing a backup and restore process can be challenging. A few months ago I developed a solution architecture deploying Neo4j into Kubernetes as a casual cluster. There’s a Medium post by Neo4j’s David Allen to explain what that configuration looks like here. Unfortunately Neo4j doesn’t officially support a casual cluster deployment, but there are community maintained helm charts endorsed by Neo4j that make this achieveable. For this solution I needed a simple backup and restore (nothing more) which is what I wanted to focus on here.
Nov 11, 2021
This article was published in the journal Towards Data Science. Please click here to access the article.
Aug 2, 2019
This article was published in the journal Towards Data Science. Please click here to access the article.
Aug 2, 2019