Nice video explaining the sizing concept.. But I have a following question.. We may have many transactional service methods in the application layer.. each may have different round trip calls to db.. so ideally cpu idle time on the db server for each of these transactions would be completely different... So on what basis, we can identify the foreground idle time of db server.? Is it just the average of all ?
The (weighed) average of all would be a good indicator for your connection pool size, yes. The video gives you the concepts. The real world is always more complex...
Nice video explaining the sizing concept..
But I have a following question..
We may have many transactional service methods in the application layer.. each may have different round trip calls to db.. so ideally cpu idle time on the db server for each of these transactions would be completely different...
So on what basis, we can identify the foreground idle time of db server.? Is it just the average of all ?
The (weighed) average of all would be a good indicator for your connection pool size, yes. The video gives you the concepts. The real world is always more complex...
First !