Build a useful investigation brief
Map concurrency through the request path
Align arrival, admission, handler start, worker or thread availability, connection pools, database work, dependency waits, response completion, and queue depth. Compare resource use at steady and burst levels. Record where measurements are unavailable and avoid inferring saturation from a correlated metric alone.
Trace overload and recovery behavior
Inspect status codes, timeout classes, queue age, retry signals, partial writes, duplicate attempts, and stale reads during the burst and after it ends. Identify which layer rejects or delays work and whether recovery drains safely. Preserve route mix and request shapes for the reproduction.
Test one bounded capacity change
Evaluate one intervention such as limiting admission, reducing per-request work, changing queue consumption, or protecting a shared connection limit. Replay steady and burst cases at the same scale. Check success, rejection messaging, ordering, duplicate prevention, and recovery time before accepting higher throughput.
What to carry forward
Deliver an evidence-backed brief with the controlled load shape, first saturated boundary, overload behavior, recovery evidence, and one experiment. State remaining capacity uncertainty. Investigation is complete when implementation can target one shared limit and verification can replay both burst behavior and post-burst correctness.
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