- 20 Sep 2023
- 3 Minutes to read
- Updated on 20 Sep 2023
- 3 Minutes to read
Device groups and individual devices currently in an active maintenance window are clearly identified in dashboards. Instead of any status indicators, the device will show as "in maintenance" and display the maintenance icon (wrench) next to its name.
(See Maintenance Window Management for information about creating and scheduling maintenance windows.)
There are three distinct ways that a managed device may enter a maintenance window:
- By being part of a selection that enters a normally scheduled maintenance window.
- By manually placing it into a maintenance window.
- By adding it to a device group that is currently in an active maintenance window.
Any downtime for a device that occurs within a maintenance window is subtracted from the total possible uptime for that device. Therefore, a device that was down for four hours in one week, but only within maintenance windows, would still show 100% uptime for that week. However, if a device is still down when a maintenance window ends, any currently active alarms will again be allowed to open incidents as normal.
Types of Maintenance Windows
One-time Maintenance Window
A one-time maintenance window is a scheduled maintenance window that only executes once and then is removed from the schedule. This is useful for planned outages that are not part of regularly scheduled maintenance.
Recurring Maintenance Window
A recurring maintenance window is a maintenance window that is scheduled to execute once, at a specified time, every day, week or month. This is useful for planned outages that are part of regularly scheduled maintenance.
When configuring a recurring maintenance window, it may be set to "enforced." This means that any other devices that are placed into the device group selected for maintenance while this window is in progress will be immediately placed into maintenance with the rest of that group. (This does not apply to maintenance windows for WebART, EmailART or service checks.) Additionally, maintenance windows using the enforce feature cannot be closed early. Take this into consideration when creating your windows.
Maintenance Window Scheduling
Users with the Power User, Admin or SuperAdmin access level may create scheduled, one-time and recurring maintenance windows on the Maintenance Window Schedules page.
To view any currently active maintenance windows, select Utilities > Maintenance Windows from the main menu to navigate to the Active Maintenance Windows page. (This page only shows maintenance windows that are currently running. Scheduled maintenance windows are not shown here. For that, you will need to navigate to the Maintenance Window Schedules page by selecting the Maintenance Window Administration button on the Active Maintenance Windows page.)
When scheduling maintenance windows, be sure not to create overlapping maintenance windows that contain the same device (or device groups), as this can cause unpredictable behavior.
Maintenance windows may additionally be created and managed through the Maintenance Window API.
Incidents and Alert Notifications
The purpose of maintenance windows are to prevent alert notifications from being sent by Netreo about devices that are offline or otherwise non-functional due to issues known and planned for by a user.
While a managed device is in a maintenance window, any new alarms generated by that device are prevented from opening new incidents (this means that no alert notifications are sent, since there is no incident to initiate them).
If any alarms are still active after the device leaves the maintenance window, those alarms are then allowed to open new incidents, as appropriate. Those new incidents will then send alert notifications and re-notifications as normal.
Existing Open Incidents
When a managed device is placed into a maintenance window, if there are any currently open incidents for that device, those incidents remain in the state that they were in when the device entered the window. The alarm(s) for the monitoring check(s) that opened those incidents will continue to update according to the schedule configured in each check, but no action groups (including ones with alert notifications) will be run -- nor will any re-notifications be sent -- while the device remains in a maintenance window.
Once a device leaves a maintenance window, any existing open incidents are updated based on any current alarm statuses. If an alarm still exists (because the issue causing it still exists), the related incident will remain open (and re-notifications resumed, if applicable). Conversely, if all alarms for an incident have cleared, the incident will be closed in accordance with normal incident behavior and the proper alert notification sent. (See Incident for more information on incident behavior.)