Getting rate limit from Split API

[davidji99]

Getting rate limit from Split API

[bulju]

Hi @davidji99, I just created a PR -> #111
Let me know if you like the philosophy that I took and if you believe that something is missing, I can rewrite it if it's needed!
As I don't have a sandbox environment I wasn't able to test it through the API with all the functionalities, I just ran a terraform plan for environments and now it's working.
But this only covers the "get" use case, not post, patch, etc.

[jramosf]

The HTTP GET use case should be good enough since the rate limiting happens when the statefile is being refreshed in the plan phase.

Only in the case of updating/recreating a very large Split environment with the provider we may hit the limit if the request happen in parallel.

[bulju]

I'll work on this and raise a PR @davidji99

[davidji99]

Hmm, this is interesting. 20 resources at the default parallelization of 10 for terraform seems quite low for a rate limit.

I'm curious, what happens if you set the -parallelism flag to say 5 (terraform apply -parallelism 5)?

In any event, let me raise this issue to my Split contacts.

[bulju]

I did try adding -parallelism=1 and the issue persists.

Based on their documentation, when you receive a 429 HTTP code (which is the rate limit), they return in their headers how much time you should wait to get a new quota. I guess we could add that to the client for doing a retry logic.

Here there is another provider which had the same issue: okta/terraform-provider-okta#186

Let me know what your contacts say about this issue.

I'm planning to work on this by the end of this week, to have a PR and then we can decide if we merge or not :)