Daniel Strobusch

> The way RPCs are canceled in gRPC leverages RST_STREAM directly. So when the client wants to cancel an RPC, it cancels its context, the gRPC client library sees this...

I tried that code in one of my projects. Unfortunately it also won't address resources retiring `Multi` responses (see also #44898). So I guess there is some alternative implementation needed...

@joe-elliott, thanks for your feedback. > 100%. I think this is a great idea and would love a PR. See #3964 – it took a few days to get it...

The grpc team is considering to add the cancel cause, see https://github.com/grpc/grpc-go/issues/7541#issuecomment-2301665760.

Found another issue, which would need to be fixed to get the cancel cause reported. [batch.go](https://github.com/grafana/dskit/blob/main/ring/batch.go#L199) would need bo the changed: ```diff case

@joe-elliott I created https://github.com/grafana/dskit/issues/576 and https://github.com/grafana/dskit/pull/577 in [dskit](https://github.com/grafana/dskit). But I am not sure if there is a process to encourage someone to look at it.

remove stale. Open tasks: * bump gRPC in case https://github.com/grpc/grpc-go/issues/7541 gets implemented * bump dskit to a recent version with cancel cause reporting * update tempo code to set a...

@joe-elliott It looks like the gRPC team is not convinced by the use of context cancel causes. If you or someone from the Tempo team wishes, you might want to...