Last week, we noticed unusual activity with regards to our pull stats of one of our images: lsioarmhf/nginx
It was getting about one pull per second (way more than usual, as it is a relatively new armhf image, not too many existing users yet). As you know, the stats count not only successful downloads, but also pull attempts that don’t result in downloads (image is already up to date). Naturally, we suspected a systemd type script continuously attempting pulls at a rapid rate.
We do have one addon for a specific OS (LibreELEC) that automates updating the images and creating the containers through the use of systemd, which we thought could be the culprit, but it is properly rate limited (limited to 3 pulls per 24 hours)
We had to take action and moved the image from lsioarmhf/nginx to lsioarmhf/nginx-armhf to put a stop to it.
However, we are wondering if the issue was the result of a bug on our end, or if it was a single user creating a faulty start script that went into a loop of docker pulls and the user simply didn’t realize it.
Is there anyone at Docker we can get in touch with to discuss this issue?