Closed KrystianMarek closed 3 years ago
Let me get with the RT team to find out what that actually means
Can I have the terraform file you used for this? I really want to run it locally.
I need to know the version of RT you're running and the TF file you're executing. It looks like you're working with an already configured instance of RT. Just because the TF provide returns an error from RT doesn't mean the TF provider is the problem.
If I can't get a copy-and-paste reproducible case, I will need to close this ticket.
We are also seeing this quite regularly, our solution was actually to limit the parallel option on Terraform to like 2 or 3. It is also more common the bigger amount of changes there are in the apply. It seems like a problem with the API not being able to keep up with the requests or some Async issues.
I would not consider it an "problem" with the API - There is no obligation for an API to support full ACID - frequently, for reasons of performance, async processing is desired and occurs. Sometimes it's a function of a data storage choice basically dictating this behavior.
A good solution would be for the go code in the provider to actually implement retry/backoff.
An enhancement request would be that the provider support high concurrency interactions with RT.
After updating an existing permission, single line update, I end up with this: