Dependency latency
Recreate poor network conditions by delaying all network traffic to a dependency by 100ms.
Latency
Dependencies
10 minutes
How this Scenario works
This Scenario adds 100ms delay to all network traffic to simulate slow dependency responses.
Why run this Scenario?
Dependencies can be slow to respond to requests, no matter where they're located. However, your services still needs to be responsive, or your customers will think your application is slow. This Scenario tests how your service behaves when its dependencies are slow to respond.
- Verify retry logic works without overloading dependencies.
- Identify bottlenecks in upstream or downstream services.
- Assess end-user impact of increased dependency latency.
- Ensure your application can still meet its SLA despite slow dependencies.
What to expect when you run it
When a dependency takes 100ms longer to respond, your service continues to meet its SLA, retries without overwhelming the dependency, and users notice little or no difference.
More Scenarios to try
Avoid downtime. Use Gremlin to turn failure into resilience.
Gremlin empowers you to proactively root out failure before it causes downtime. See how you can harness chaos to build resilient systems by requesting a demo of Gremlin.
.svg)