Bug: timescale_service resource not propagating its attributes

[pieeri]

Hi Timescale team,

We're currently deploying TS services thanks to your Terraform provider (v1.9.0) and are eager to integrate your provider with our multi tenant client deployments.

We have managed to programmatically create a Timescale VPC and accept automatically the Timescale peering connection within our AWS cloud, while launching a service and associating it with an internal DNS record (Route53).

While this works admirably, we reset the instance password through the Timescale web UI and ran into a situation where the password provided by the timescale_service resource didn't refresh, despite resync'ing / refreshing our Terraform plan.

To allow our application to connect, we had to destroy such a Timescale service and recreate a new one from scratch (and we ran into #177 as well, where Terraform couldn't recognise that it was removed).

Please let us know if you need more information to investigate the issue. Many thanks in advance!

[Khyme]

Thank you for your feedback.
I'm investigate as I know we labeled password a sensitive field and terraform internals might have a specific dynamic for it that we did not implement yet.

[fatima-kscope]

@Khyme I'm storing the timescale password in a secrets manager when terraform is run. It becomes null in the secrets manager everytime I run terraform, most likely because of this issue. Also, weirdly, only seeing this in one deployment. Is there a workaround/fix in progress?