Description
Unfortunately I don't have much info on this one, but thought it still worth noting, in case someone else has a similar issue and this may help?
I did a quick bit of googling, and apparently, the slowlog is for queries that take too long (or longer then they probably should). This is a total stab in the dark, but it appears that a common cause of random "hangs" is the MySQL query cache.
I'm not neccessarily recommending that we do it by default, but perhaps it's worth considering modifying TKLBAM to (optionally) flush the query cache before making a backup?! It may have an overall negative effect on the longer term MySQL response but if it makes backups more reliable, then perhaps it's worth it?
FWIW, I discovered someone reporting that their tests suggest that the DB size sweet spot for the query cache is between ~64MB -> ~128MB.