That's all true. We've increased the timeout for now, however, the proper fix requires quite some effort. We need to move the GridFS blobs to Swift and then storage management will become simpler (production wise). This also covers multi-charm store syncing at the same time. I'm also going to propose that we introduce quotas for charm store on user namespace. You can ask for more space but that's how we can prevent from some CI to upload gigabytes of data per day by mistake. Log analysis will help as well, but that's also on ToDo ATM :-/ On Sun, Jul 10, 2016 at 9:28 AM, John A Meinel