Handle 429 from auto-model service

[charisk]

Shows a slightly more helpful message when we receive 429s from the auto-model endpoint.

Checklist

N/A - internal only:

• CHANGELOG.md has been updated to incorporate all user visible changes made by this pull request.
• Issues have been created for any UI or other user-facing changes made by this pull request.
• [Maintainers only] If this pull request makes user-facing changes that require documentation changes, open a corresponding docs pull request in the github/codeql repo and add the ready-for-doc-review label there.

[koesie10]

LGTM, although the progress doesn't get cleared when I run into a rate limit error. I think you'll need to add await this.clearProgress(); just before return; in the generateModeledMethodsFromLlm method.

[charisk]

LGTM, although the progress doesn't get cleared when I run into a rate limit error. I think you'll need to add await this.clearProgress(); just before return; in the generateModeledMethodsFromLlm method.

Ah, I was contemplating whether to tidy this up in this PR. I've added this.clearProgress() in the catch block to make sure we clear on other failures. too.

[koesie10]

Ah, I was contemplating whether to tidy this up in this PR. I've added this.clearProgress() in the catch block to make sure we clear on other failures. too.

Thanks! We might want to move that into a catch block inside the generateModeledMethodsFromLlm method to also catch errors when the query fails, but this is error is much more likely to happen, so that can happen in a follow-up.