You have right, we have tested it and we are not getting any results.
We strongly believe that after the last update, something has been changed and now the alert is not working.
Please note that before the update the alert was working normally without any issues.
Mempool data will be filled/updated with polling/discovery results
maybe you reduce filter options to get it working? thinking about mempool_index
Donât know what it is for
to verify âWe strongly believe that after the last update, something has been changed and now the alert is not working.â would it be possible to roll back updates till youâve got it working again and post here the first commit in which it doesnât work anymore?
i recommend to do this on a testing system.
Resurrecting this as keen to get a solution too - to the problem of Virtual memory usage on linuxâŚ
I think this query might be close, adjusting as necessary to return results (to sanity check) the 40% and 10%
MariaDB [librenms]> select device_id,mempool_descr,mempool_perc FROM mempools where (mempool_descr = âVirtual memoryâ AND mempool_perc > 40) OR (mempool_descr = âSwap spaceâ AND mempool_perc >= 10) GROUP BY device_id HAVING COUNT(*) = 2;
You need a question mark in every advanced query as a placeholder for the device_id.
Also, why is this an advanced query? and why do you need both virtual and swap? I would suggest only checking swap. virtual memory is just physical memory + swap.
Alternatively, make two alerts, one for swap and one for virtual if you really want to monitor virtual.