…ately after it is used. This will address the flock() issue: #123
This is handy to know if there are more than LIMIT queries to display, or if you're looking at the full list
It would be useful to visualize just how much each of the top queries contribute to the aggregated values on the graph. We already have support to pivot based on hostnames, adding a pivot based on checksums would take a little more work. The major change is that the current search page can fire ajax requests for both the table and graph result data simultaneously because they have no particular dependencies on each other. To get the top N checksums we need the data from the table result first. This change will have to enable that -- when this checkbox is checked, we'll delay getting the graph results until the table results are returned, then use them to get the list of checksum values. Once we have those we can send the right data to the graph ajax request to make this happen.
Previous behavior was to force the timezone to be America/Los_Angeles when none was set. This caused issues because the data in the table didn't agree with the offsets in the interface and you would get incorrect query results Now the default is to use the system timezone, which produces more reasonable results.
there was some work done on this in a previous commit, but this update fixes another minor issue which was preventing the correct max column width from being used.
previously it was using a URL to download an RPM which was in the LATEST directory on the percona server which means the specific version there would change by using a specific version we'll have to update this occassionally but it's more reliable for now
also fixed some permissions issues with the vagrant boostrap.sh setup process. I moved the cron job from root to vagrant user, but forgot about file permissions on the slow log file. Moved it back to root.
…es are different (fixes #104)
I'll update the wiki asap!