Describe the bug
As already reported in the past at https://github.com/TykTechnologies/tyk/issues/2248, the health check currently always reports 200 even if the actual state is either warn or fail. This causes issues as the pod would try to accept requests but fail as it isn't fully working.
On a on premise setup, we noticed that some of the pods spinning up failed the initial fetching from the control plane. Instead of being restarted, they started accepting requests which raised the number of 404 errors. A functioning health check would have avoided this.
Reproduction steps
Steps to reproduce the behavior:
Remove redis/Setup RPC with wrong configurations
Actual behavior
The container is still healthy and receiving requests
Expected behavior
The container should be unhealthy and restart until healthy (if ever)
Version and Environment
Describe the bug As already reported in the past at https://github.com/TykTechnologies/tyk/issues/2248, the health check currently always reports 200 even if the actual state is either
warn
orfail
. This causes issues as the pod would try to accept requests but fail as it isn't fully working.On a on premise setup, we noticed that some of the pods spinning up failed the initial fetching from the control plane. Instead of being restarted, they started accepting requests which raised the number of 404 errors. A functioning health check would have avoided this.
Reproduction steps Steps to reproduce the behavior:
Actual behavior The container is still healthy and receiving requests
Expected behavior The container should be unhealthy and restart until healthy (if ever)