Add verbose rate limit handling and improve API error handling

[dnorth98]

Summary

• Adds visible feedback when the GitHub API rate limits requests (previously the script would silently hang)
• Prints message to stderr showing sleep duration and resume time when rate limited
• Adds is_rate_limited() helper to detect rate limit responses from API
• Adds --output-file option to write results to a file
• Improves error handling for unexpected API responses throughout the script

Test plan

• Run script until rate limited and verify message appears: Rate limited by GitHub API. Sleeping for X seconds until HH:MM:SS...
• Verify script continues normally after rate limit period expires
• Test --output-file option writes results correctly

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR adds verbose rate limit handling to provide feedback when the GitHub API rate limits requests, and improves error handling throughout the script. It also adds a --output-file option to write results to a file.

Changes:

• Added make_verbose_rate_limit_handler() to patch the GitHub client and print rate limit messages to stderr
• Added is_rate_limited() helper function to detect rate limit responses
• Added --output-file option and output() function to write results to a file
• Added error handling checks for rate limiting, non-2xx status codes, and unexpected response formats in get_workflow_runs()
• Added rate limit checks at various API call points throughout the script

Comments suppressed due to low confidence (2)

get-deployment-metrics.py:507

• The file writing at line 507 uses "\n".join(output_lines) which will not include a trailing newline, and each message in output_lines is a complete line without a newline character. This means the output file will be missing the final newline that's typically expected in text files. Consider adding f.write("\n") after the join, or using f.write("\n".join(output_lines) + "\n") to ensure the file ends with a newline character.
  get-deployment-metrics.py:372
• After checking for rate limiting, there's no check for other API errors (non-2xx status codes) or unexpected response format before accessing workflow_durations as a dict at line 369. If the API returns an error status, workflow_durations might not be a dict, which could cause an exception. Consider adding status code validation and type checking similar to what was added in get_workflow_runs().

                                "WARNING: GitHub API rate limit exceeded while fetching timing data. Results may be incomplete."
                            )
                            break

                        # Some jobs may not have run at all
                        if "run_duration_ms" in workflow_durations:
                            job_duration = workflow_durations["run_duration_ms"]
                        else:
                            job_duration = 0

                        workflow_total_duration += job_duration

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.