Discussion about wellness endpoint

[antoinesd]

We planned to introduce a new endpoint /wellness to solve issues like #130 . What are the pros and cons for this introduction?

[xstefank]

Just to clarify, it would be /health/wellness right?

[antoinesd]

Yes and introduction of @Wellness qualifier

[hyperman1]

My interest in this is for a situation where my application is up, but a required dependency (e.g. upstream web service) is down.

Quarkus on openshift uses live and health to restart a container if they are not healthy. If my application is OK but a dependency is down, automatically restarting my application wont help. Also, I might be partially up but running degraded. Monitoring this so we can warn the owner of the dependency is still relevant.

[xstefank]

@hyperman1 hi, thank you for showing interest in this feature. We've already implemented it in our downstream SmallRye project to test out the API before we can include it in the specification (it should be available for you if you are using Quarkus). Would you be able to test it out and give us any feedback, please?

[hyperman1]

@xstefank That's in fact where I found it, after the quarkus team suggested a custom health group. It does perfectly what I want:

• Appears in /health (custom health groups don't and @health is deprecated )
• Does not appear in /health/live or /health/ready.

I get nervous about using @experimental code in production, so I added the comment above in the hope of getting it standardized. Thanks for your extremely fast reaction.