Liang Mei

Reduce aggressive retries in case of service busy error

1

## Expected Behavior When there's resource exhausted error, the retry between client/server and within history service is too aggressive. This amplifies the load on the server and often makes the...

**Is your feature request related to a problem? Please describe.** Currently when autoforwarding is on for all apis, it would be hard to investigate standby cluster/workflow issues (because all requests...

Add rpc request metrics with retry in consideration

1

**Is your feature request related to a problem? Please describe.** Right now we have rpc metrics for single requests that does not factor in retry options. Since client does retries...

Check https://github.com/temporalio/go-samples/issues/30

**Is your feature request related to a problem? Please describe.** Delayed callback after starting workflow will not take effect. Cadence lib should prevent this from happening. Repro: send signal in...

It's a common requirement that various workflow threads want to use lock/mutex to protect critical sections. Right now user needs to implement this logic with future or await. We should...

https://github.com/uber/cadence/issues/2327

See original discussion https://github.com/uber/cadence/issues/2891

Prometheus does not allow dash in method names.