The Same Error Message, Three Completely Different Times
A dashcam-to-YouTube upload pipeline failed with the exact same quota error on three separate debugging passes — and each time it was a genuinely different bug wearing the same symptom, including one caused by the previous fix itself.
This article is also available in Traditional Chinese: 中文版 — same content, just a language difference.
A batch upload job failed with a YouTube quota-exceeded error. The obvious assumption — today's daily quota had been hit again — was wrong, and disproving it opened up a chain of three unrelated failures stacked on top of each other, the last of which was accidentally introduced by the fix for the second.
Don't trust a plausible-sounding error without checking its evidence
The pipeline uploads recorded dashcam footage to YouTube via a relay running on a separate machine, reached over a reverse SSH tunnel from the laptop doing the uploading. A batch of uploads failed immediately with a quota-exceeded error — a completely reasonable thing to believe at face value, since that's a real, common failure mode for this API. But a quick check of the actual evidence undercut it immediately: the log file recording upload progress was over a week stale, not from the current day, meaning whatever produced this error hadn't actually attempted a fresh upload recently at all. A quota error can't be about today's traffic if nothing from today ever reached the API.
Unlock this article to keep reading, or subscribe for unlimited access to everything. See Pricing for details.