This is a project that I’ve created with the goal of visualizing metrics from my zettelkasten in my Grafana instance deployed in my Homelab. It was the first service that I’ve written that was deployed to a Kubernetes cluster.
I’ve written it in Go, since the language is an excellent fit for small services and I was feeling like writing it.
The initial idea was to expose the metrics on a Prometheus scrape endpoint, but I learned that the pull model was not a good fit for registering historical data. With that in mind, I’ve migrated to InfluxDB for the initial release.
Things I learned so far:
- Prometheus has a great documentation on writing exporters
- Parse markdown with goldmark
- Better test assertions with testify
- Using the
fs.FSinterface instead of directly reading system files withosimproves testability. Got that one from learn Go with tests - Prometheus is really not the best tool for this job, I’ve switched to InfluxDB
- The flux query language looks a lot like Elixir’s pipe syntax, and it’s quite good at creating pipelines for transforming data
- goreleaser is a great tool for creating GitHub releases
- Working with GitHub’s container registry is quite easy with GitHub Actions