[Bug] Race Condition in hashicorp/go-metrics PrometheusSink

[fridrik01]

I identified a race condition in the Cosmos SDK's telemetry dependency of hashicorp/go-metrics. The bug is in PrometheusSink when multiple goroutines create the same metric concurrently, metric updates can be lost.

I've pushed a fix for this and created a PR in the upstream repo with tests demonstrating the issue and verifying the solution.

However, I haven't seen any updates on that PR, pointing to hashicorp/go-metrics having little maintenance activity. Could the Cosmos SDK team help get this merged upstream? This affects all Cosmos SDK applications using Prometheus metrics and can cause incorrect metric values in production environments.

[linear]

SDK-419