Configuring Tyk Liveness Probe in Kubernetes

Product FAQ
, , ,
1

Hi,

I am currently facing a challenge with configuring the liveness probe for Tyk in our Kubernetes environment. According to the Tyk documentation here, the /hello endpoint is recommended for health checks. However, I have observed that this endpoint always returns a 200 OK response, even when Tyk is not fully operational due to the unavailability of certain components, such as Redis.

Given this behavior, I am seeking guidance on best practices for configuring the liveness probe in Kubernetes to ensure it accurately reflects the health of the Tyk gateway, especially when critical components are down.

I would appreciate any insights or recommendations you may have.

Thank you for your assistance.

2

Hi @Uri_Parush,

Thank you for your question.

If you define a mock response endpoint then it would only return a 200 when the APIs are successfully loaded; a nice way to check both that the gateway is running and that the APIs are loaded.

Or you may set up a Virtual endpoint, which could make a couple of extra requests to other components and return 200 if all is fine. For the Redis situation in particular, some extra tool like Webdis may be needed.

Hope this is okay.