Currently some deployments fail/timeout (mainly hydra) as we get caught with crashloopbackoff containers, rolling update expects things to be healthy before proceeding. Hoping if we force update these containers it won't hang around and proceed. This could potentially be dangerous if we do it for every deployment, so it should be a flag we pass into the CLI at deploytime.
Description
Currently some deployments fail/timeout (mainly hydra) as we get caught with crashloopbackoff containers, rolling update expects things to be healthy before proceeding. Hoping if we force update these containers it won't hang around and proceed. This could potentially be dangerous if we do it for every deployment, so it should be a flag we pass into the CLI at deploytime.
Parts of the app this will impact
Possible Implementation
Technically how could this be implemented. Which areas of the code could it touch.
We just need to pass in
forcebased on k8s client docs https://github.com/kubernetes-client/python/blob/549482d8842a47c8b51122422e879e7cc497bf88/kubernetes/docs/AppsV1Api.md#patch_namespaced_deployment Requires QAYes / No
Design Work
Any links or notes that relate to design work for this feature
Backend Work
Any information about BE such as request URL's, JSON responses, errors or PR's
Acceptance Criteria
Clearly defined criteria for exactly how the feature should work
When I do this It should do this
Then if I do that It should do that
Additional Information
Related Links