Josh Imhoff

Thank you! Can we get a 22.1 backport? CC serverless is running 22.1. Major upgrade can take some time in practice. The lack of these HBs has led to inactionable...

Ah sorry for confusion! Major upgrades don't lead to unavailability. The issue fixed by this PR has nothing to do with major upgrades. CC serverless is simply running 22.1, and...

Some convo with @sumeerbhola, @JuanLeon1, and @tbg: Sumeer says: > I just read https://github.com/cockroachdb/cockroach/issues/71169. This is typically hard, and I don’t know any storage/db service in google that managed to...

> I'd say that the most fruitful thing we could do would be measure the KV request latencies and try to pay attention to see whether they're slower than expected....

Returning to this ahead of 22.2 planning. The idea is to measure E2E latency and error rate while excluding latency and errors caused by workload choices. For example, latency in...

@andy-kimball makes a great point that we don't need to do this perfectly for it be very useful, def for cloud & prob for on-prem. - Concretely, we can select...

> But what if we implemented this with an allowlist that permitted configuration of which apps to emit bucketed metrics for? I think could be useful! OTOH I don't see...

I'll summarize for @rytaft since this ticket is long. We want latency metrics that track what our customers are experiencing E2E, as E2E behavior is what users actually care about...

Anyone have a rough idea of how much eng work this one is? Obviously scope is not decided yet, and so we have some flexibility. Just looking for a rough...

Hi! I'm an SRE at CRL. The SRE team does lots of k8s stuff. If you want to do this, I can review the PR. Are you doing it because...