performance: Augment existing Indexer/Conduit metrics & logs
winder opened this issue · comments
Make sure the metrics and logs available from Conduit are sufficient for measuring performance.
At a minimum, Conduit needs:
- the total transaction count per block added to the log.
- the # of transactions per type (axfer, pay, app call..)
- the time it took to ingest a block.
Some of this is already available in Indexer.
(out of scope) the idea is that we will use some tool/script to analyze the logs. This is separate from the Prometheus data (general system performance), which uses the metrics endpoint (which reports averages).
AFAIK this is out of date.