att-comdev / promenade

This project has moved to OpenStack.
https://www.airshipit.org/
Apache License 2.0
11 stars 15 forks source link

Promenade Lifecycle Managed Components missing Liveness and Readiness Probes #48

Open jezogwza opened 6 years ago

jezogwza commented 6 years ago

Is this a bug report or feature request? (choose one): ENHANCEMENT - Missing Resilience Feature

Python Version (output of python --version):

Development or Deployment Environment?: Development

Release Tag or Master:

Expected Behavior: All of the components deployed by promenade should be resilient, and able to partake or leverage the Kubernetes resilience features. Specifically Kubernetes should be able to identify when one of these components is not functioning properly, even when the pod appears to be running. In general this is achieved using health Probes. This is a request to make sure all the promenade deployed components include health probes where appropriate.

What Actually Happened: One of the kubernetes-apiserver pod wasn't responding. Upon killing the pod and allowing kubernetes to re spawn it it started to work again. With a heath probe that should have occurred by itself.

How to Reproduce the Issue (as minimally as possible): Not sure how we got to that situation. The behavior was easy to spot. kubectl would stagnate every other request, necessitating a CTRL^C , after which reissuing the command kubectl would usually work.

Any Additional Comments:

mark-burnett commented 6 years ago

Thanks for filing this issue @jezogwza

Adding probes to all the charts is definitely valuable.