Hard dependencies have a timeout
Every synchronous call the service cannot work without has a declared timeout.
CriticalChecked automatically
- Applies to:
- Services with hard synchronous dependencies
- Rule id:
- resilience.hard-deps-have-timeouts
Why it matters
A call without a timeout waits as long as the other side takes, which during an outage is forever. Each waiting call holds a thread or a connection. When the pool is empty, your service stops answering too, including the requests that never needed the slow dependency. This is how one slow service becomes a platform outage.
How to apply it
- Set a timeout on every outbound call: connection and response.
- Derive it from the dependency's normal latency (a little above its p99), not from a round number.
- Keep it shorter than your own callers' timeout, so you fail before they give up on you.
- Decide what happens when it fires: an error to the caller, a fallback, or a queued retry.
Example
checkout → payments: connect 100 ms, response 800 ms (payments p99 is 450 ms). Checkout's own callers time out at 2 s.
References
Checklists with this rule
Related agent skill: resilience-patterns