[eli-r-ph]

💡 Motivation and Context

The v1 SDKs now retry HTTP 404 and 429 on capture, like 408 and 5xx, so a temporary failure from capture or a proxy in front of it doesn't drop the batch on the first response.

• _RETRYABLE_STATUSES adds 404 and 429; they leave _TERMINAL_STATUSES. They use the same backoff, jitter, Retry-After handling and retry budget, and raise CaptureError with their status after the last attempt.
• The retry and exhaustion tests now run for 404, 429 and 503. The terminal-status test uses 402 in place of 429.
• Migration guide: the retried status list now includes 404 and 429, and says how it differs from 7.x. A changeset describes the change.

💚 How did you test it?

• pytest posthog/test (5338 passed) and ruff check/ruff format. The new 404 and 429 cases fail with the old status set.

📝 Checklist

• I reviewed the submitted code.
• I added tests to verify the changes.
• I updated the docs if needed.
• No breaking change or entry added to the changelog.

If releasing new changes

• Ran sampo add to generate a changeset file

🤖 Agent context

Autonomy: Human-driven (agent-assisted)

Authored with Claude Code (Claude Opus 5.5), directed by @eli-r-ph.

🤖 Generated with Claude Code