Reputation: 106
I run an application that usually triggers a heavy update every 2 hours, by queuing heavy tasks, using the cron mecanism. This has been working well for months.
However, the december 16th of 2012, this url has been called (by user-agent AppEngine-Google) every few seconds between 15:17 and 15:51, launching hundreds of heavy tasks. This resulted in an explosion of my quotas and forced me to switch to the paid version of the application, in order for my website to stay alive.
Anybody having the same issue? Any idea of what happened and how I could avoid this problem in the future?
Upvotes: 0
Views: 210
Reputation: 1
I had the same issue.
I don't have an answer, but I think this is a task queue problem.
I have 4 cron jobs and some task queue tasks piled up. Everything was normal until 14:05Z (16th 6:05 PST?).
At 14:06Z and 14:07Z, two of my cron jobs were called (at their scheduled times) and finished with 200. After that, AppEngine-Google started to call the same jobs few times per every minutes. The disorder vanished after 14:50Z and no issues right now.
During the period, one of my task queue tasks was called at 14:11Z and finished with 503 (this was an expected failure). The task was scheduled to retry some hours later, but was called hundreds of times in an hour. The task's retry count was not incremented.
My guess is that something went wrong in task queues ("__cron" and, for me, "default") and the tasks were not removed until 14:50Z.
My app's App ID is vidssage.
Upvotes: 0