Fix example/grafana-integration

[angristan]

Which problem is this PR solving?

Currently, the grafana-integration example doesn't work properly: if you run docker-compose up in that folder, services will start but only logging will work, the metrics and tracing won't.

Description of the changes

• Fix traces endpoint config for hotrod (traces export: Post "http://localhost:4318/v1/traces": dial tcp 127.0.0.1:4318: connect: connection refused)
• Fix hotrod metrics scraping (endpoint has moved to the frontend service)
• Pin Jaeger/hotrod tags to prevent future issues
• Fix Grafana dashboard (metric names, labels, migrate to new time series panel)
• Add default Grafana credentials to README
• Fix the loki container being stuck on shutdown by setting shorter timeouts (bug with the driver: if loki is not reachable and loki-docker-driver is activated, containers apps stops and cannot be stopped/killed grafana/loki#2361 (comment))
• Update Grafana/Loki/Prom pinned versions (especially to get Grafana UI improvements)

How was this change tested?

docker-compose up 🙂

Checklist

• I have read https://github.com/jaegertracing/jaeger/blob/master/CONTRIBUTING_GUIDELINES.md
• I have signed all commits
• I have added unit tests for the new functionality
• I have run lint and test steps successfully
  • for jaeger: make lint test
  • for jaeger-ui: yarn lint and yarn test

[yurishkuro]

Thanks for the fixes.

Could you elaborate on the value you're getting from this example? It seems 99% about Grafana, Jaeger almost not needed or used. I am tempted to delete this example altogether.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (1dd4b60) 95.62% compared to head (96e5315) 95.59%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4980      +/-   ##
==========================================
- Coverage   95.62%   95.59%   -0.03%     
==========================================
  Files         319      319              
  Lines       18760    18760              
==========================================
- Hits        17939    17934       -5     
- Misses        659      663       +4     
- Partials      162      163       +1     

Flag	Coverage Δ
cassandra-3.x	25.64% <ø> (ø)
cassandra-4.x	25.64% <ø> (ø)
elasticsearch-5.x	19.90% <ø> (-0.02%)	⬇️
elasticsearch-6.x	19.90% <ø> (-0.02%)	⬇️
elasticsearch-7.x	20.05% <ø> (ø)
elasticsearch-8.x	20.12% <ø> (-0.02%)	⬇️
grpc-badger	19.53% <ø> (-0.02%)	⬇️
kafka	14.07% <ø> (ø)
opensearch-1.x	20.03% <ø> (-0.02%)	⬇️
opensearch-2.x	20.05% <ø> (ø)
unittests	93.35% <ø> (-0.03%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[angristan]

Sure :) Well, I would agree Prometheus + Loki might not be relevant in this context indeed. We could probably delete them from the example to keep it leaner.
However, the value I got from it was to see how Grafana could use Jaeger as a data source, I'm pretty pleased with the result.

I'm happy to rework the MR and remove Loki and Prometheus, but feel free to close it if you want to delete it altogether 👍

[yurishkuro]

@angristan do you know if it's possible to configure Grafana to open Jaeger UI for traces instead of the built-in view? I don't mind metrics & logs, it's the trace view that bug me, because it effectively shows "this is how you don't need to use Jaeger" (even though it's still used for storage, but not the UI)