timeout
Enforces a maximum end-to-end latency from request receipt to response headers.
On this page
Enforces a maximum end-to-end latency from request receipt to response headers.
Configuration Notes
This does not cancel the upstream connection; the upstream has already responded by the time this check runs. It is useful for enforcing SLAs. Because enforcement is post-hoc, the request has already been delivered upstream (non-idempotent side effects have happened), and a hung upstream is not bounded by this filter; configure cluster timeouts for that.
Configuration
| Field | Type | Required | Description |
|---|---|---|---|
timeout_ms | integer | yes | Maximum allowed elapsed time from request receipt to response headers, in milliseconds. Requests that exceed this limit receive a 504. Enforcement is post-hoc: the check runs once response headers arrive, so the request has already reached the upstream and any non-idempotent side effects have already happened. It does not bound a hung upstream; use cluster connection/read timeouts for that. |
Example
filter: timeout
timeout_ms: 5000 # 5 seconds