My team and I recently switched a relatively large MongoDB deployment (0.5T originally) onto AWS DocumentDB. Strangely, DocumentDB is taking up way more resources than I think it should. On our old cluster we never went over 1500 IOPS, but now we're well over 50k in AWS, which is terrifically expensive. Also, the storage size has ballooned to 2T - almost as if the data were no longer compressed at all (WiredTiger would have been compressing things in the old cluster).
What could be happening here? As far as I know there were no code changes. Are there any tools I could use to figure out what the IOPS are being used for? Are there other unique sources of I/O usage that I might not be thinking of? A >10x jump seems totally crazy to me.