Coverage

“The agent is online” does not say that a system is being watched. The Coverage page shows eleven components per system and says for each one whether the data is arriving. A green square is in order. Any other square is a place where monsys is not watching right now, with the reason.

When all data arrives but the eight setup steps are not done yet, the system shows Not set up instead of green.

The components

Component	What it is	In order when
Heartbeat	The agent checks in	last check-in within 5 minutes
Load	Processor, memory, network	last reading within 10 minutes
Disks	Usage per disk	last reading within 10 minutes
Containers and services	What runs and what stopped	last report within 15 minutes
Inventory	Software, packages, ports	last scan within 13 hours
Updates	Pending security updates	last check within 26 hours
Certificates	Expiry dates on the system	last check within 48 hours
Logins	Successful and failed logins	data in the last 7 days
Backups	Check for a recent backup	set up and checked within 26 hours
Reachability	Outside check	set up and checked within 30 minutes
Agent version	Runs the latest version	equal to the published version

What the squares mean

Status	Meaning
In order	The data is arriving
Limited	It works, but slower than it could (older agent)
Stopped	The data was arriving and has stopped
Never received	Nothing has ever arrived about this for this system
Error	The agent tries to collect it and gets an error
Still starting	The system is new, the first result is expected
Not set up	Has to be set up by you first (backups, reachability)
Not applicable	Does not exist on this kind of system

The verdict per system is Fully watched, With gaps or Silent. Silent means the heartbeat is missing: no other square can be trusted then.

Containers and services

monsys remembers what runs on a system. The rule is simple:

• What ran stably for half an hour is expected to keep running.
• If it is not running for more than three minutes, monsys opens a problem: “Container web is not running” or “Service nginx.service is not running”. A restart or a new rollout fits inside those three minutes.
• A service in the failed state is always a problem, also when monsys never saw it run.
• Once it runs again for two minutes, the problem closes by itself.

Nobody has to write a rule for this. If you want it differently, you choose per container or service:

Choice	Effect
Must run	Always a problem when it is not running
Auto	monsys decides by the rule above
Not required	Never a problem, for example for a job that runs now and then

Containers started with docker run --rm are never expected: they are meant to disappear. If you close a problem by hand while the container is still down, monsys learns that container again instead of reopening the problem at once.

A system that is offline does not get a problem per container. The silence of the system itself is the problem then.

Older agent

An agent from before 10 November 2026 does not yet report every two minutes what is running. monsys then sees the containers through the inventory, every six hours, and shows the component as Limited. Services are watched once the agent is updated. That happens by itself when automatic updates are on.

Archiving

A system that no longer exists should not stay in the list as silent forever. Open the silent system on the Coverage page, give a reason and choose Archive.

• The system no longer counts as silent.
• The open problems of that system are closed.
• The history is kept.
• If the system reports again, it comes back by itself.

A system that reported in the last 15 minutes cannot be archived. monsys archives systems that have been silent for 30 days by itself.

API

Call	Purpose
GET /api/v1/coverage	All systems with their components
GET /api/v1/agents/{id}/workloads	Containers and services of one system
POST /api/v1/workloads/{id}/expected	{"expected": true}, false or null for auto
POST /api/v1/agents/{id}/archive	{"reason": "..."}
POST /api/v1/agents/{id}/unarchive	Restore