hey, we seem to have gotten pulumi into a flaky st...
# general
r
hey, we seem to have gotten pulumi into a flaky state after importing manually created aws resources. It rarely deploys, replaces unchanged things, fails updates etc. Repeatedly re-running pulumi sometimes works. Any pointers? 🙂 It's a ts awsx classic project that's been running flawlessly basically unchanged for years. Diff in thread
only diff I can see in
pulumi stack export --file
(other than reordering items) is:
urn:pulumi:prod::lesslie-backend::awsx:ecs:FargateService$aws:ecs/service:Service::backend
->
outputs
->
networkConfiguration
->
subnets
is sometimes an empty array and sometimes contains two subnets. The corresponding input subnet id array is empty.
l
Are you running refreshes? If you are, are there differences between code and AWS?