Resiliency

Fault injection, request mirroring, rate limits, retries, and timeouts.

📄️Fault injection

Inject artificial latency into requests to test how your clients and services handle slow responses.

Send copies of requests to alternative backends for shadow testing.

Enforce budget and spend limits per key by controlling request and token usage.

Configure automatic retry attempts for failed backend requests.

Set request and backend timeouts to prevent long-running requests.