Policies
One policy tree for alert rules, monitors and tickets across the workspace, customers, sites and devices, and how a device's policy is worked out.
Updated
Policies (in the RMM group of the sidebar) decide what RevoRMM watches on each device: the alert rules, the monitors and where alerts become tickets.
Policies and levels
A policy is a named set of settings, such as "Standard", "Servers" or "Front desk PCs". It can be used in many places. A policy sets only what you choose in it; everything else comes from above.
You assign a policy to a level:
- the workspace: every customer, including customers you add later;
- a customer;
- a site of a customer;
- a single device.
How a device's policy is worked out:
- Closest wins. For each alert rule and for tickets, the level nearest the device that sets it wins. A server site's policy can raise the disk threshold while the workspace's still applies to everything else.
- Monitors add up. A device gets the monitors of every level above it.
- Switch off what you inherit. A level can switch off a monitor it inherits, for itself and everything below it (open the level, then turn the monitor off under Inherited monitors).
Where they apply shows the tree: the workspace, each customer, its sites and any device with its own setting. Click a level to change its policy and see what applies now, with the level each setting comes from. A device's Monitoring tab has the same view under Policy. To give many customers the same policy, tick them and click Set policy of ….
A policy that is assigned somewhere can't be deleted: unassign it first, so no level loses its settings by surprise. Copy starts a new policy from an existing one.
Monitors
Monitors run on the device (RevoAgent 0.11 or later), except script monitors, which RevoRMM runs as approved scripts.
| Monitor | Checks | Can fix |
|---|---|---|
| Service | A Windows service, systemd unit or launchd job is running (or stopped) | Start it (or stop it), at most every 15 minutes |
| Process | A program is running, or a program that shouldn't be isn't | End the unwanted program, at most every 15 minutes |
| Event log | New Windows event log entries by log, source, event ID, level and text | — |
| System journal | New Linux journal entries from the units you list, at a priority or worse | — |
| Ping | The device can reach a host | — |
| TCP port | The device can open a port on a host | — |
| Website | A web address answers from the device with an expected status and text; warns before its certificate expires | — |
| Script | An approved script exits with 0 (anything else fails) | — |
Each monitor checks every 1 minute to 1 day (script monitors: every 15 minutes or less often). Its alert settings are the severity, how long it must fail before an alert opens (or, for event monitors, how long without new events before the alert resolves), whether it resolves by itself, and a re-arm time.
Presets start an event-log monitor for blue screens, unexpected shutdowns, application crashes, disk errors, Windows Backup failures and UPS events (APC and Eaton software), and journal monitors for Linux UPS daemons and service errors. Adjust them after adding.
A device runs at most 64 monitors; beyond that, the effective view marks the ones left out. A monitor for another operating system (an event log on Linux) shows as "not on this system". Devices on RevoAgent 0.10 keep CPU, memory, disk and offline rules and run script monitors; the other monitors start when they update to 0.11.
Script monitors
A script monitor runs a script version that was approved with a passkey, with the arguments you set, on each device the policy reaches, while the device is online. Exit code 0 passes; any other exit code fails and its first line of output becomes the alert's message.
Because a script monitor runs on every device its policy reaches, setting one up, and anything that starts script monitors on more devices (assigning such a policy, switching one back on), needs a person signed in with the rights to run and to manage scripts (owners and admins; never an API token), within the number of devices your role may run a script on at once. From 10 devices you type the policy's name to confirm. The devices count towards the hourly limit that pauses scripts on unusual activity, together with scripts people run; the monitor's later runs don't. Pausing scripts (Settings → Scripts) stops script monitors too. Versions with secret parameters can't be used in a monitor.
Patch policies
Patch policies follow the same levels. A customer's and a single device's patch policy are chosen in Patching; a site's is chosen here (open the site, then Patch policy; it needs the patching permission). The effective view shows which patch policy a device gets and from which level.