On 3 November the workspace was slow to list files for about twenty minutes after a cache server restarted cold. The API itself was fine; only the browser listing dragged while the cache warmed. We now warm the cache before an instance takes traffic.
A short fix: file names with accented or non-Latin characters were being written to the export manifest in a normalised form that did not always match the file inside the zip. They now match byte for byte. Existing exports are unaffected; re-export if you want a clean manifest.
Routine firmware updates on the storage volumes ran overnight across zones. Uploads were briefly slower in a couple of zones during each window. Nothing to do on your side.
On 2 October last year a bad release made share pages return errors for eighteen minutes. We never wrote it up properly, so here it is, late: a template change shipped without the share route's data, every /s/ link 500'd, we rolled back in eighteen minutes. Accounts and files were untouched. We added a check that a release serves a real share page before it goes out.
On 4 March, 02:00-03:00 UTC, we will upgrade the account database in every zone. Sign-in will blink out for a few minutes per zone. Uploads already running will not be affected. We will post here when it is done.
On 29 March uploads in Amsterdam were slow for about half an hour while a storage volume rebalanced. Nothing was lost, but it was slow enough to notice, and a few of you did. The rebalance should have run overnight; a scheduling bug ran it at 07:40. We moved the job back to its window and added an alert if it ever starts during the day again.
Folder deletion had been leaving a few orphaned file records behind in rare cases. No files were affected; the records pointed at nothing. We fixed the cause and swept the leftovers tonight. If your file count looked one or two higher than it should have, it is correct now.