Stream: infra-ops

Topic: Dataverse Observability & Monitoring

Oliver Bertuch (Oct 01 2026 at 15:30):

Folks here might be interested in the introduction and small tech demo I gave about using OpenTelemetry with Dataverse during today's CT WG meeting. Here's the recording: https://drive.google.com/file/d/13zEC-lzhOWD5fPxXvKoJNlaq7l0MVkI2/view?usp=sharing

If anyone is interested in tinkering around and push this further, please reply here and get the ball rollin'. :roller_coaster: Note: this is not at all limited to containerized installations! Also, it's a cross cutting topic, as devs should benefit from tests with traces and performance testing!

Thanks @Akio Sone @Ash Manda @Philip Durbin 🚀 @Balázs Pataki for asking so many questions and listening to my crazy ideas!

I also created an issue to be able to reference in a draft PR: #12762. Here's the code I was able to figure out so far: #12763. Please note: you need to build custom base and application images locally to play with this!

Oliver Bertuch (Oct 01 2026 at 16:22):

I'd be happy to talk about experimenting with this. Possible topics:

• Use OTel agent instead of builtin. Would it give us more out of the box?
• Why only "/api/v1" traces - what's going on with the http.route. Is it our code or Payara screwing up?
• Logging with JUL may be a suboptimal choice going forward. Refactor? Logging bridges? Logging util to include context and make Payara use JSON logs?
• Solr / Postgres OTel: hand over span from Dataverse?
• JSF integration
• React SPA integration
• JUnit, Cypress, ... integration
• Add spans to subsystem borders in the codebase?
• Playing with spans + metrics, e.g. can we monitor the indexing behavior with metrics per span?

And so much more!

Tony Nguyen (Oct 02 2026 at 15:36):

@Oliver Bertuch Adding OTel will be very helpful. There are 2 opinions:

• If we can request for a datadog opensource version will be great - especially https://www.datadoghq.com/product/real-user-monitoring/ if available for open-source - https://opensource.datadoghq.com/.
• https://github.com/signoz/signoz this is one of the deployment option.

Oliver Bertuch (Oct 02 2026 at 15:39):

@Tony Nguyen thanks for the suggestions! @Ash Manda suggested Grafana Cloud as well. I've read good things about https://openobserve.ai/. Before we can start using something like this for tests we run, there's probably a long way ahead to integrate all of these things... :wink:

Philip Durbin 🚀 (Oct 02 2026 at 15:41):

our work is never done :smile:

Ash Manda (Oct 02 2026 at 15:50):

Philip Durbin 🚀 (Oct 02 2026 at 15:52):

ha, a classic :rofl:

Oliver Bertuch (Oct 02 2026 at 15:53):

Classic, but I wouldn't dare calling any of the observability cloud offerings a "standard" :wink:

Philip Durbin 🚀 (Oct 02 2026 at 15:57):

Is OpenTelemetry the first one?

Oliver Bertuch (Oct 02 2026 at 16:00):

Pretty much I think. At least on https://opentelemetry.io/ they dare to call themselves a standard. Seeing how it is a CNCF backed project, I think it's fair to say it is a "defacto standard". (It's no RFC, ISO, ... AFAIK)

Philip Durbin 🚀 (Oct 02 2026 at 16:02):

back in my day people wrote RFCs

Last updated: Nov 02 2026 at 07:01 UTC