Possibility to replace long-running pods / lifecycle policies

[FoxDawg]

Hi there,

is there any possibility to create a trigger that automatically recycles long-running instances?
Like, for example, if I want my pods to be recreated after 2days to keep the cluster fresh and healthy - is there any such thing with KEDA?
Thanks!

[zroubalik]

Hi, KEDA doesn't support this at the moment.

[FoxDawg]

Thanks!
Any other chance then to tell KEDA to take down the oldest instances in case of downscaling events? That would be a decent compromise in our environment.

[zroubalik]

The actuall scaling is handled by Kubernetes HPAs, KEDA doesn't have any saying in it.

[FoxDawg]

Good point. Maybe a cron job in Kubernetes might be a way to go. Thx!