IssueTracker

Problem you have encountered:
- The lack of retry-on-stale connections logic has resulted in Cloud Run latencies.

Investigation and workarounds:
- The API clients are globally scoped in the code, they get cached and are [reused between invocations](https://cloud.google.com/run/docs/container-contract#outbound-reset). However, this results in issues when the connections are not kept alive for a determined amount of time.

- According to the [documentation](https://cloud.google.com/run/docs/troubleshooting#connection-reset-by-peer), this can be solved by enabling the keepalive in each socket, however, this depends in the library being used.

- The main workaround to solve this, besides mentioned above, is to manually implement a retry strategy in the application's code.

What you expected to happen:
- The backend of the library should implement the logic necessary so it doesn't need to be implemented from every environment that makes use of it instead.

Steps to reproduce:
- Build an application in Cloud Run based on this library.
- Do not configure keepalives or retry logic.
- Await for outbond connections to reset.