Keep the lesson for the next incident
Document transport ownership
Record pool size, wait policy, response consumption, release on success and failure, timeout, cancellation, retry, and restart behavior. Define which callers share capacity and who owns a concurrency change. Record the caller that owns each connection.
Keep lifecycle fixtures
Retain fast, concurrent, slow, failed, cancelled, retry, and restart cases with expected counters and local states. Include the original leak or pressure shape so transport upgrades can be compared without real payloads.
Review capacity signals
Watch active and waiting connections, release mismatches, timeouts, retries, response duration, and caller queue age by client revision. Assign an owner and threshold. Close the follow-up only when shared callers exercise the lifecycle checks.
What to carry forward
Close learning with transport rules, lifecycle fixtures, caller ownership, and pressure signals. Keep upstream limits visible. The durable outcome is a finite, inspectable response to pool pressure rather than an unmeasured capacity increase. Review lifecycle counters after transport changes.
Technical background: GitHub documentation.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow