we are experiencing performance issues due to our less than desirable architecture of bundling OLTP and DSS. we are working to separete them, but in the mean time, i want to be able to proactive prevent run away adhoc queries
from dessimating our production OLTP instance.
in looking at the profiles, i believe we want to start with trying
cpu_per_call. what i don't get is how cpu time relates to real time. to your understanding what is the relationship between the time column in solrais top utlity and the cpu_per_call time in seconds.
for example, i see a process now, 12341 that has 34:57 time in top, what does that equate to in cpu_per_call time?