Block submission metrics

[avalonche]

📝 Summary

Migrating the metrics to OpenTelemetry and set up more detailed profiling to measure the request latencies of block submission and proposer api.

⛱ Motivation and Context

Improve observability and latency issues in boost relay

📚 References

---

✅ I have run these commands

• make lint
• make test-race
• go mod tidy
• I have seen and agree to CONTRIBUTING.md

[0x416e746f6e]

remark:

not so relevant if we just drop this counter, but generally this approach is not desirable as each new combination of labels (slot and floorSecIntoSlot in this case) will cause prometheus to start a fresh counter (which will eventually explode as the size of the scrape payload will keep growing with time).

[avalonche]

how do we label metrics not using attributes then?

[0x416e746f6e]

there will be labels added automatically by the scraper (things like instance, instance_name, and so on).

important is to keep all possible label permutations limited.

e.g. in the current case floorSecIntoSlot would be still fine (and maybe interesting to capture, to be able to see the the breakdown of requests per second => and there would only be 12 possible labels for that.

but the slot would change every 12s => therefore infinite count of labels. unless we somehow make the scraper "forget" old labels, so that scrape request size would be limited in size.

[0x416e746f6e]

let's goooo