10:38 a.m.
For simplicity of keeping all aware of the conversation, let's
use this
thread and start discussing what and how to bring OpenTelemetry
experimental features to Operate First users, and how to begin delivering
Op1st environment value back to OTel as an environment user/contributor.
I'll reply to this message to take a part of the thread into the "what does
this look like for everyone" discussion.
Sally should hop in anytime and correct or expand on anything here, most of which came out
of our discussion.
From the tracking issue https://github.com/operate-first/community/issues/83:
Vision includes:
- The project will run latest upstream code with experimental features enabled.
- Open source devs can tap into the experimental features, e.g. OpenTelemetry Tracing
enabled etcd, then use the API server to tie that to their own code so they can get an
end-to-end trace for the time of a request (for example.)
- The Op1st community cloud SRE team also benefits from having the latest in metrics,
logs, and tracing capabilities. OTel Collector is able to take exported data and export to
different platforms, such as Prometheus.
- Roadmap should include having a way to use GitOps to provide granular control over which
experimental feature is enabled for which environment. Then people can test against a
subset or all of the experimental features.
- What can we expose to SRE/Data Scientist users? Can we let them have some level of
access to what OTel is providing to services running on Op1st CC?
- What value is the OpenTelemetry project able to get?
- Demo environment so show people how it works in a running Kubernetes environment.
- Whatever operational data it can collect with the various workloads it tracks in the
project