Stat for "Value too large"

There has been few cases that has taken some time to debug where the application was sending documents that were too large to the cluster and as a result was using up all of the network bandwidth. To help debug this problems quicker in the future, I think it would be useful to have a stat to track "Value too large" responses. We already have an stat for get_misses (Key not found) and NMVB responses.

I understand that the application logs will have the exception/log however the users who raise support cases rarely have access to the application logs and take a long time to upload.

In the future it would also be useful to have the key and hostname of client that did the operation that failed available to the Admin in the UI so they can deal with it themselves.

While this is similar to MB-12332 I believe the fact that this is an error makes it different enough to require it's own MB.