d10r
commented
1 year ago Changes implemented.
A reverse proxy was added to the pre-existing prometheus push gateway, providing us with a nice URL.
The identifier was solved with a random uuid. Since prometheus uses a system of multiple labels per metric, we can leave that persistent - multiple instances per instances are well covered due to the addition of chain_id as label.
Instead of converting the health report, a custom report is generated for telemetry. That's because we don't need that much data, and because that structure isn't nicely mappable to prometheus format despite GPT claiming so.
see https://github.com/superfluid-finance/superfluid-sentinel/pull/186#pullrequestreview-1629946257