Add retries for "worker env WorkerEnvId(workerenv-XXXXX) not found"

[mgyucht]

Changes

Due to the eventually consistent nature of worker environment creation as a part of workspace setup, certain API requests can fail when made right after workspace creation. We have some exception handling for these in the SDKs, but a new case has appeared: "worker env WorkerEnvId(workerenv-XXXXX) not found" (see databricks/terraform-provider-databricks#3452).

This PR addresses this issue. Furthermore, it moves the transient error messages into autogeneration so that there is a single source of truth that applies to all SDKs.

One other small change: I removed running unit tests from the autogeneration flow. It removes a small amount of convenience (users need to run make test after regenerating the SDK), but it speeds up the devloop when iterating on code generation, and it allows the release flow to continue to make a PR before failing. This PR can be modified as needed to fix any test failures or compilation failures, as per usual.

Tests

Added a unit test to cover this.

• make test passing
• make fmt applied
• relevant integration tests applied

[jiropardo]

Hello @mgyucht @tanmay-db , does this take into account that XXXXX changes depending on the worker env ? I may have missed this initially in #3452