Wesley Kim

@wisre we are in the process of trying out what we think are fixes here, should have a release cut next week sometime!

@tpinellia is there an issue with using v1.29.1? I believe we have some dependencies we haven't sorted out that still require that version.

@shrek thanks for the contribution! The approach seems to make sense, can you create a pull request and flesh out the test cases? Thank you!

@naughtyGitCat is the raw prom metric data equivalent to the raw m3db data? Could you post a dump of datapoints from both raw timeseries (prom vs m3db)? We can try...

@skupjoe where did you get the base coordinator config? We can look into updating that. Also pull requests are welcome if you'd like to contribute a change!

@acostama is this implemented?

+1 to this approach, and `max_shard_step_ms` satisfies our use cases as well. I wonder if we could detail the use cases for `max_shard_size_ms` / `max_shard_size_points`, or else if it's worth...

@jranson gotcha I may be confusing myself, but how would `max_shard_size_points` help in this case, given the query could load a variable number of timeseries? Also one thought re: naming....

One more comment: if we do split queries up, I imagine https://github.com/tricksterproxy/trickster/issues/291 may become more of a frequent issue? We noticed in the [missing range](https://github.com/tricksterproxy/trickster/blob/main/pkg/proxy/engines/deltaproxycache.go#L314-L330) handling that errors seem to...

> if the delta proxy request to retrieve the remaining 1% fails, the originating client request should be returned a full failure This is the desired behavior, yeah. The natural...