Wen Long

Hello, is this issue friendly to newcomer? If it is, maybe I can help on this.

@kakkoyun regarding the latency, currently from the metrics page, I saw there is a `grpc_client_handling_seconds` histogram. Is this already represent the BatchWriteClient latency? Metrics example: ``` grpc_client_handling_seconds_bucket{grpc_method="WriteRaw",grpc_service="parca.profilestore.v1alpha1.ProfileStoreService",grpc_type="unary",le="0.005"} 1 ```

This PR is based on some of the comment from https://github.com/prometheus/exporter-toolkit/pull/70

@roidelapluie hello, I got two questions: 1. May I know what do you mean by make the parameter more generic? 2. Regarding the skip TLS cert validation, you mean when...

> The use case for this is that you can serve /-/health without TLS authentication but the rest with authentication because kube probes do not support certs Got it! Let...

Some information sharing. By default, when controller received SIGTERM, it will directly issue a SIGQUIT to nginx process without waiting(https://github.com/kubernetes/ingress-nginx/blob/main/internal/ingress/controller/nginx.go#L373). I set --shutdown-grace-period to 15 and the downtime seems to...

Hello, I don't have much experience on writing concurrent code, might make some mistake. Thanks for guidance in advance. :D

@kakkoyun thanks for the review and guidance! yes I need some help. I am not sure how would it works if we removed the `batchDone` channel. If we removed `batchDone`,...

I think this already fixed by https://github.com/parca-dev/parca/pull/1396?

I am working on it.