Add alert for an increase in seq scans

This adds an alert that fires if a host executes 20% more sequential table scans compared to the hour before.

Some questions we need to answer before merging:

1. Do we want to increase the time range (e.g. 6 hours)?
2. Did I write the correct query?
3. What should the for option be set to? I set it to 15 minutes, but should this perhaps be set to 1 hour as well?

cc @bjk-gitlab