[Lifecycle] Do not fail deployments

[thisthat]

Goal

Help user prepare their pre-checks without blocking any deployment

Details

Users that want to adapt Keptn for pre-checks might want to test that they nail down the right set of tasks and evaluations before they can safely run them in production. To support them, we can add a new config option that allows Keptn to run them without blocking the deployment.

Acceptance Criteria

• Add a new entry block-deployment in KeptnConfig
• Default value should be true
• If block-deployment is true, then the status of WorkloadInstance for pre-tasks and -evaluations is based on the respective action results
• If block-deployment is false, then the status of WorkloadInstance for pre-tasks and -evaluations is always set to succeed at the end of their execution

DoD

• Documentation is updated
• A user can execute pre-checks without blocking deployments if the configuration is enabled

[agardnerIT]

This is great and will help potential users understand that Keptn is "safe by default". To that end, I would advocate for the default being false.

[YashPimple]

@thisthat would take up this issue

[thisthat]

Hey @YashPimple thanks 👍

[agardnerIT]

I'm confused. This is assigned to @YashPimple but Yash says @thisthat will take it up. Concerned that this important safe default is falling through the cracks.

I'm happy to work on this if it needs people - just point me in the right direction of the code please and assign to me.

This is on the critical path for adoption.

[thisthat]

Hey @agardnerIT, @YashPimple will work on this and by the end of the week he will provide a PR :)

[mowies]

any updates on this @YashPimple ?

[agardnerIT]

I've just been hit by something like this again.

In this case, I'm using ArgoCD to roll out my cluster in steps (sync waves).

Keptn rolls out in wave 5. Backstage (and other components) rolls out after that in wave 6.

Keptn could not start (specifically the certificate operator) due to cluster sizing (insufficient CPU).

But now that Keptn components cannot start, the admission webhook endpoints are not available. Everything that comes after this is completely blocked.

PLEASE PLEASE can we make Keptn safe by default!

[odubajDT]

Since there was not coding activity on this issue since Dec 11th, setting the ticket again as free for assignment. Please feel free to re-apply for assignment if you want to work on the ticket!