Drift and job watch
Finds the slow slide, and the scheduled job that quietly stopped.
Some failures never spike. A machine gradually uses more memory each week, or a nightly backup stops running and nothing complains because nothing is watching for absence.
Drift watch keeps a four-week behavioural baseline for every machine and compares the present against its own past, not against a threshold somebody guessed at during setup.
Job watch handles the opposite problem: something that should have happened and did not. It checks every 15 seconds and raises an incident for a job that has missed its window — the backup nobody noticed had been failing since Tuesday.
At a glance
- drift baseline
- 4 weeks per machine
- job checks
- every 15 seconds
- detects
- gradual change and absence
- compares against
- its own history
All six engines run on every plan with agents.
Install on one machine and the first score lands 20 seconds later.
Deploy an agent→