Caching error responses and letting subsequent requests to pass through for the TTL period greatly increases the average performance.
With my current infra setup, I'm seeing around ~5ms on average when the nginx module needs to make a request to iprepd. When the error response (specifically 404's) is cached, the average latency overhead is around 0.5ms.
Caching error responses and letting subsequent requests to pass through for the TTL period greatly increases the average performance.
With my current infra setup, I'm seeing around ~5ms on average when the nginx module needs to make a request to iprepd. When the error response (specifically 404's) is cached, the average latency overhead is around 0.5ms.