Maintenance windows and mutes

Maintenance windows, maintenance mode and mutes all stop alerts being raised for a while, and a planned state says that something is a test machine, unused or retired, so its alerts are judged that way. None of them loses anything: the rules keep running the whole time, and when the window, maintenance or mute ends, anything still wrong is raised within a minute.

An alert that is already open when a window or mute starts stays open.

The maintenance windows page#

Open Alerts and click the Maintenance windows tab at the top of the page. It lists every window in three groups: Active now, Upcoming and Past. Each row shows where it applies (a device, a whole site or a policy), the customer, the schedule, who made it and who ended it early. Filter by Scope, Customer and Made by.

When the work finishes early, click End now on an active window. That occurrence ends at once and anything still wrong is raised within a minute; a repeating window still comes round again next time.

Add a maintenance window#

1. Go to Alerts, open the Maintenance windows tab and click Add window.
2. Under For, choose A site, A device or A policy, then pick it. A site window covers every device at the site, including ones added while the window is on.
3. Fill in the name, how it repeats and the times, as below, and click Add window.

Add a maintenance window to a policy#

1. Open the policy in Settings > Monitoring.
2. Under Maintenance windows, click Add window.
3. Give it a Name, for example Patching.
4. Choose how it Repeats: Once, Every day or Weekly (then tick the days).
5. Set when it starts and ends, and the Time zone. Times are on that clock, so a 22:00 window stays at 22:00 when the clocks change.
6. Check the summary in words at the bottom, for example Once, 26 Sep 22:00 to 27 Sep 00:00 (Europe/London), and click Add window.

The window applies to every device the policy reaches. For example, a customer policy might have a Sunday patching window, and a Reception PCs group a nightly one.

Tip: Line your monitoring windows up with the maintenance windows in your patch policies, so restarts after patching do not page anyone.

Add a maintenance window to one device#

1. Open the device.
2. On the Monitoring card in the side column, click Add beside Maintenance windows.
3. Fill it in as above.

While a device is in a window, its health shows a maintenance chip and a notice. The Monitoring card lists the device's own windows, its site's and its policies', with the next one or when the current one ends, and End now. A site's page lists and adds the site's windows on its Alerting card.

Maintenance mode#

Maintenance mode is for work happening now. While a device is in maintenance, no module raises alerts or pages for it (monitoring, patching, backup, security, printing and infrastructure), automation rules skip it unless they are set to include devices in maintenance, its page shows a banner with End now, and lists and search show a wrench beside it.

1. On the device page, open the menu in the header and choose Maintenance mode. You can also use the row menu or bulk bar in the device list, the customer and site menus, or Cmd+K.
2. Choose how long, or until when, and say why.
3. Click Start maintenance.

It ends by itself at the time you chose. Patch policies can also put devices into maintenance mode while installs and restarts run, ending when the device is back online. Time in maintenance is left out of the uptime figures.

Planned states: test, unused or retired#

Some things are not broken, they are planned that way: a lab pair you test upgrades on, a server that is switched off on purpose, a site that has closed. Planned state... in the menu of a device, site, customer, server check, firewall or controller, NAS or virtual desktop deployment says so, and alerts from every module about it follow what you set.

1. Open the record's menu and choose Planned state....
2. Under Mark it as, choose:
   • Test or lab: alerts about it stay visible in the list but are never ticketed or sent to anyone.
   • Unused: alerts about it are closed as planned, because nothing is expected to work.
   • Retired: alerts about it are closed as planned, because it is no longer in service.
   • In maintenance (on records that have no maintenance mode of their own): alerts are held until the window ends, and what is still wrong then is told after a short grace.
3. Optionally set Until and a Reason, then click Set.

The dialog lists what is set now, with End on each. Ending a state judges its alerts again at once. A device that is retired in Devices > Lifecycle counts as retired here too.

Automatic fixes wait for maintenance#

Maintenance is also when Tenvara may put things right by itself. An alert type set to Automatic in the maintenance window in Settings > Alert remediation runs its fix only while the device's patch window is open or the device is in maintenance mode, never at a random moment in the working day. See What raised an alert, and fixing it from the alert.

Mute an alert#

You can mute from three places:

• a monitoring alert: Mute mutes that rule on that device;
• a device page: Mute on the Monitoring card, or press M;
• a check's history: Mute at the top.

1. Under What to mute, pick one check or rule on the device, Offline, or Every monitoring alert on this device.
2. Choose For how long: 1 hour, 4 hours, 1 day, 3 days, 1 week, or Until... a date and time.
3. Optionally say Why. The reason is shown to the team beside the mute.
4. Click Mute.

See and end mutes#

Settings > Monitoring lists everything Muted now, with the device, what is muted, until when, who muted it and why.

• Click Unmute to end a mute early. Anything still wrong is raised at the next minute.
• A device's Monitoring card lists its own mutes, each with Unmute.

Note: Maintenance windows and maintenance mode hold back alerts from every module: monitoring, patching, backup, infrastructure and the rest. A mute holds back one monitoring rule only. Security detections are never held: an attack on a machine in maintenance is still an attack. To quieten an alert from any source for a while, snooze it: see Alerts. To stop a whole kind of alert, overall or for one customer, turn its type off: see Alert types.