Skip to content

Implement liveness probe

Alexander Chueshev requested to merge add-liveness-prob into main

This MR implements a liveness probe for the model-gateway service. The liveness probe is available via the /monitoring/healthz endpoint.

To successfully pass the liveness probe, the server must have access to Triton. If the server is unable to reach Triton, the endpoint returns a 503 error

Please, check the model-gateway deployed for testing purposes only - https://console.cloud.google.com/kubernetes/deployment/us-central1-c/ai-assist/fauxpilot/model-gateway-test/logs?project=unreview-poc-390200e5.

Ref #14 (closed)

Edited by Alexander Chueshev

Merge request reports