on 2015 Mar 10 9:55 AM
Request level logging can be very useful in seeing what is actually happening in a live database, but the higher levels of logging (eg SQL, PLAN etc) do involve a fair degree of server load - especially with many hundred simultaneous users. Ideally one could just log a single connection but the way the docs read:
Output to the request log can be filtered to include only requests from a specific connection or from a specific database, using the sa_server_option system procedure.
everything (ie for all connections / databases) is being collected (with the consequent server load) but the output is being filtered down to a single user.
Am I reading too much into the word "filtered" or is that the correct interpretation?
If my understanding of the docs is correct, you should be able to collect only RLL data for a specified connection via
CALL sa_server_option( 'RequestFilterConn', connection-id );
To cite:
RequestFilterConn
Filter the request logging information so that only information for a particular connection is logged. This can help reduce the size of the request log file when monitoring a database server with many active connections or multiple databases.
So in my humble opinion that would mean the data collecion is already filtered, not the later RLL output. - Note, that's just another WAG for today:)
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
It isn't absolutely clear is it? The word "logged" could be taken as meaning "only data for a particular connection is written to the log" rather than "only data for a particular connection is collected"
It talks about keeping down the log file size but doesn't mention the server load.
Mmmm.. I may be being paranoid but it doesn't mean they aren't out to get me!
User | Count |
---|---|
71 | |
11 | |
11 | |
10 | |
9 | |
9 | |
7 | |
6 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.