This ☞ Quora.com thread is listing the following largest MongoDB deployments:
- ServerDensity (BoxedIce guys)
- Squarespace & Stickybits
While being aware that we are obsessed with numbers, I find the original question too generic. I’d probably formulate it completely different:
- what are the largest collection in terms of number of records? (nb: the size of each document is also important)
- what are the largest collection in terms of size on disk? (nb: having 100mil records with 32bytes size is completely different from having 1mil records with 4MB size)
- what are the largest collections in terms of index sizes? (nb going to disk can be painfully slow — read about the Foursquare outage and post-mortem analysis — and that MongoDB will use all existing memory on a machine)
- what is the largest deployment in terms of running servers? (nb: partition characteristics (for example Yahoo! PNUTS/Sherpa is caping each shard at 500MB) and machine characteristics are also important)
- what problems have you hit for each of these?
Now that would probably give a lot more information about how to proceed on your next project. But at the end of the day, you’ll still need to try it out yourself.