Database flooded with thousands of requests while not in use
Posted: Wed Apr 14, 2021 4:54 amFor the second day now, our vicidial agent conversations have been crippled with poor audio and disconnects.
A concerning flag we can see is that the dedicated ViciDB is pegged at 100% CPU (according to the load indicator, curiously top shows a much more conservative value hovering around ~3.5%).
In running a mysql general_log trace, we observe literally thousands of hits despite it being off hours and there are no agents are logged in, no calls happening.
Seems we probably have some over-zealous tasks running or configured. Any advise where to look/tune?
Server is a recent Dell performance model, I don't have specific specs but it's an SSD, with plenty of multi-core CPUs and RAM.
A concerning flag we can see is that the dedicated ViciDB is pegged at 100% CPU (according to the load indicator, curiously top shows a much more conservative value hovering around ~3.5%).
In running a mysql general_log trace, we observe literally thousands of hits despite it being off hours and there are no agents are logged in, no calls happening.
Seems we probably have some over-zealous tasks running or configured. Any advise where to look/tune?
Server is a recent Dell performance model, I don't have specific specs but it's an SSD, with plenty of multi-core CPUs and RAM.