I use matomo and queued tracking for the tracking of an ecommerce platform. I regularly encounter that there is a problem with tracking data and the queued tracking processing breaks, leaving the data in redis to get bigger and bigger and bigger until all available ressources are taken and redis gets killed by OOM.
My solution so far was to move data from one queue (the on with the broken data) into another queue and process it from there.
On the other hand it should actually be possible to just delete, drop or pop the malformed entry. would that entry be on the left or right side of the queue?
Thanks for your help,
P.S. redis lpop and rpop both did not result in successfully removing the broken dataset