Bump high disk usage alert to 80%
instance down: reduce threshold to 2 minutes
This conveys the intention a bit better.
Fix regexes for disk device identification
These are regexes, not globs.
meta_rules: switch to delta as metric
The previous version of these rules was incorrectly testing for >5
events per second, which isn't the desired behavior. By measuring the
delta, we determine if the number of events is >5 over a 10 minute
.build.yml: corrections per package rename
Add build manifest and auto-deploy