there's some interesting recalculation going on wi...
# aws
r
there's some interesting recalculation going on with the
containerDefinitions
property of a Fargate TaskDefinition. I've got updates on every
pulumi up
, even when I don't change anything about the container. I'm trying to fill in the deltas between the configurations that I see when doing
pulumi up --diff
to see if I can get the states to ever sync up and not require re-pushing, but there are definitely some default values that are represented in the 'deployed model' that aren't represented in the 'about to deploy' model
I did manage to get the changesets to line up, it seems to be a few missing properties in the calculated changes that cause it to get out of sync. it looks like the
healthcheck
property doesn't register the defaults correctly?
healthcheck.{interval, retries, timeout}
must all be specified exactly or else the container definition is replaced every time. Anyone else seeing this?