My linux laptop (core i5 + 16GB RAM) recently started to freeze and become unresponsive so I had to power cycle it.

This highly unusual behaviour repeated itself the next day so I looked seriously into it.
The reason it froze was the RAM and swap were filled by akonadiserver taking over 14GB.

After a fresh restart I noticed akonadiserver using 95-100% of a CPU core and slowly eating RAM until there's no more available.

akonadictl fsck process starts but does not complete.
finds a few errors: "Found 2 collections without RID." and "Found 34 items without RID." and always stays stuck after displaying "checking collection 103"

akonadictl vacuum completes correctly in a timely manner with nothing unusual to report.

.local/share/akonadi/db_data/mysql.err contains a few errors like this:
[Warning] Aborted connection 760 to db: 'akonadi' user: 'user' host: '' (Got an error reading communication packets)

I have no idea what to do next. Are there other options than akonadictl fsck and vacuum ?
how can I debug this further to identify and fix the cause of this ?

As a last resort is there a way to reset akonadi data entirely and if so what should I backup first ?

Thanks