Prometheus metrics
A good improvement would be mev-boost producing prometheus metrics with /metrics endpoint
open to a PR for this. what metrics are you thinking of?
I was thinking about metrics around validators registered and blocks produced, wdyt?
Dropped a draft PR, largely the metrics collected thus far are for more operational/performance metrics, would be good to add validator and block-level metrics additionally, would be great to hear thoughts on this.
Yes, I agree, validator and block level metrics would be quite handy Thanks!
Going for a sidecar approach for this, as opposed to the "deep" integration originally proposed in #294 given comments from @metachris
These changes are deep and extensive into the core system and logic. Considering that mev-boost is run at 50% of all validators, it is too risky and impactful for adding a large amount of new logic in the core flows at the current time.
Perhaps consider collecting the metrics from the logs, or thinking about a possible plugin interface. 🤔
Latest is on: #370 The idea is that you'd run a grok-exporter with an mev-boost flavoured configuration Some care points:
- Changes to log format would affect the metrics exported
- Mitigation: Config owned/managed/maintained in the mev-boost repo (would ensure alignment?)
- The fstab/grok_exporter project is not currently maintained afaics
- Mitigation: it's a pretty simple (but powerful) go tool, i'm happy to contribute if any critical bugs come about, this monitoring is non-invasive, so you can always fall back to logs (or some other grok implementation if required)
- There is currently no maintained docker container for grok-exporter