# Timeout / Slow Dependency

## Symptoms
- Requests fail with `timeout` or `deadline exceeded`
- Upstream latency spikes in metrics
- Retries triggered with backoff

## Checks
- Measure RTT and latency between caller and dependency
- Inspect dependency saturation (CPU, connections, queue depth)
- Check long-running queries or GC pauses
- Validate timeout/retry configuration and circuit breakers
- Review recent deploys or infrastructure changes impacting latency

## Fix
- Tune timeouts/retries to realistic values
- Optimize slow queries or handlers
- Add caching/batching if appropriate
- Scale dependency horizontally or vertically