I've been neglecting working with metrics and observability in general in the last couple of years but I do need those for a project now so I am looking to catch up.
I've done the reading of the
metrics crates and managed to successfully fiddle with several different configurations of
tracing -- liked it. Haven't practiced
metrics yet for the simple reason of not having a clear view about how the whole thing fits together. I would like to remove this lack of knowledge.
To my question:
Main thing I am interested in: are there good reference / example projects that demonstrate how to collect metrics in a complete manner that I can take inspiration from? By that I mean not only Rust source but also e.g.
Dockerfiles for Jaeger, Prometheus, Kafka, Grafana and what-have-you.
What I'd also love to hear from you:
- Clues about which exact pieces of libraries or external software that I'll need. Alternatives to each of them would be very welcome.
- Regularly used hosted solutions. I am not against them but I'd do some effort into trying to self-host first. It will still be very valuable to know how do people go about it with hosted solutions.
- Diagrams orienting the reader into what fits where, what's the general stack, which are the mandatory and optional pieces, some pros and cons, etc.
Apologies for the rather vague and wide question. I already have done some homework but I find myself a bit lost -- so I'll appreciate any informative takes that you might offer.
Thank you for your time.