Alan Protasio
Alan Protasio
Hi, I did not see this issue before so i opened a [new one](https://github.com/thanos-community/promql-engine/issues/105) for bottomK and topK. Those aggregations seems to need some additional work so i think it...
This seems very interesting... Besides running a single query on multiple pods this would also allow compute parts of the query closer to the data (on the storage nodes) and...
Indeed, another possibility would be to support sharding natively, so we could split aggregation over multiple pods?
I think that this change should be more explicitly called out that the memory is being reused by default on the release notes as it can break some edge cases...
cc @GiedriusS that is implementing the same in thanos.
I think make sense to change it to default! Im afraid about the breaking change indeed.. we may break customer on next release if they did not explicitly configure this?...
>I think we can start by supporting optional use of memberlist, and as for changing the default, we can transition to a later version. Cortex already support memberlist.
I think this LGTM but im afraid of making the cortex config even more complex. I wonder if those kinds of limits would not live better in another component like...
Ok.. Ship it! I think I would still try to deprecate the overlapping flags in a following PR
``` # Shard blocks across multiple store gateway instances. This option needs be set # both on the store-gateway and querier when running in microservices mode. # CLI flag: -store-gateway.sharding-enabled...