Graceful Failure

[Maxusmusti]

When a cloud provider fails to provision a machine, InstaScale should be able to gracefully handle: either by removing the failed machine and re-trying, or by scaling down the whole request if incomplete (or both, with a configurable timeout).

[Bobbins228]

@Maxusmusti I have been trying to trigger a failure phase in a machine but I have not been able to come across a way to do this. Even trying to scale machines which are not available to our org didn't work.

Have you any idea on how to trigger this scenario? Also how often would a failure like this occur?

[dimakis]

@Bobbins228 did you ever figure this out?

[Bobbins228]

I thought I had made a comment here but yeah if I remember correctly you have to create a machineSet template with an incorrect subnet attached to it which will cause a Node to enter that failed phase when machines are scaled up. That way you can replicate this behaviour.