Riya Chakraborty

metrics(indexer): Add metrics around join time for chained processing strategies

2

Currently, we have a total shutdown time for the indexer; however, we don't have a better breakdown to understand which strategies make up this shutdown time and how much.

feat(generic-metrics): Add dropped percentiles to aggregation options

2

Refactoring the way we used an older option for enabling 10s granularity so its purpose is clearer. Enable org-level and use case-level of disabling percentiles using the newly created Sentry...

Today, as DLQ is currently implemented, the DLQ buffer is never "cleared" until consumer process is shut down. As a result, consumers will keep, in-memory, multiple partitions' worth of buffer...

We currently don't have these logs, which will also be helpful for breaking down per-consumer