---
title: Maintenance windows in WatchFor
description: Suppress alerts and exclude planned work from uptime and SLA — one-time or recurring, by monitor or by tag — without pausing your checks.
canonical: https://watchfor.io/docs/monitors/maintenance
---

# Maintenance windows in WatchFor

Suppress alerts and exclude planned work from uptime and SLA — one-time or recurring, by monitor or by tag — without pausing your checks.

A **maintenance window** tells WatchFor "expect disruption here — don't page
us, and don't count it against our uptime." During the window, the monitors
you choose have their **alerts suppressed** and the time **excluded from
uptime and SLA** — but the checks **keep running**, so you can still see
whether the maintenance actually broke something.

> **Why not just pause the monitor?** Pausing stops the checks, so you get no
> data and no idea whether your change worked. A maintenance window keeps
> monitoring and simply marks the time as planned — you stay informed and
> your SLA stays honest.

## Create a window

Go to **Alerting → Maintenance → New maintenance window** and set:

- **Name** (and an optional description shown on status-page notices).
- **Monitors** — pick specific monitors and/or **tags**. A tag target covers
  every monitor carrying that tag, including ones you add later. Leave this
  empty for an [announcement-only](#announcement-only-windows) window.
- **Starts**, **Duration** (5 minutes to 30 days) and the **timezone**.
- **Repeats** — **One-time**, **Daily**, **Weekly** or **Monthly**, with an
  optional **Repeat until** date. Recurring windows keep the same local time
  across daylight-saving changes.
- **Announce on status pages** *(one-time windows)* — post a public
  "Scheduled maintenance" notice ahead of time (see [below](#announce-on-your-status-page)).

The window activates on its own when the time arrives and lifts itself when
it ends — or you can [drive it manually](#manage-a-running-window). A window
can also be **disabled** without deleting it.

## What happens during a window

- **Alerts are suppressed** for the covered monitors — no notifications
  through any channel while the window is active.
- The monitor shows a neutral **Maintenance** badge alongside its real status
  — an outage is never hidden from you.
- The maintenance time is **excluded from uptime and SLA**: it is treated as
  *neutral*, neither up nor down. Uptime bars and the SLA report call it out
  explicitly (for example *"excludes 2h of scheduled maintenance"*) — never
  silently hidden.

If a real, unexpected problem happens during the window, the incident is
still **recorded and visible** in your history — it just doesn't page you or
dent your uptime. If that problem is **still going on when the window ends**,
you are told: the incident's first notification goes out within about a
minute of the window closing (the timeline shows a *Silence ended —
notifying* entry), and it appears on your status page at once. An incident
someone already acknowledged during the window stays quiet — acknowledged
means "seen it".

A window covers a monitor that is **paused** when the window starts, too:
resume it mid-window and it inherits the suppression and the uptime
exclusion instead of paging you.

## Effect toggles

Each window has two switches, both on by default:

- **Suppress alerts** — turn off to keep alerting while still marking the
  time as maintenance (useful for a "watch closely" window).
- **Exclude from uptime / SLA** — turn off to let the time count normally.

## Announce on your status page

Maintenance is planned work, so the window is the **single source of truth**
for it — you schedule and announce it in one place instead of posting a
separate notice by hand. Status pages show what your windows publish.

Turn on **Announce on status pages** (one-time windows) and pick which
[status pages](/docs/status-pages) should carry the notice:

- A **"Scheduled maintenance"** entry appears ahead of time, with the
  window's start/end and your description.
- It lists the **affected components** — the ones whose monitors the window
  covers — so visitors know exactly what is involved.
- The notice tracks the window: **Scheduled** → **In progress** when it
  starts → **Completed** when it ends.

### Announcement-only windows

Need to tell customers about planned work on something you *don't* monitor
(a billing provider, an office move)? Create a window with **no monitors**,
turn on **Announce on status pages** and choose the pages. It posts and
tracks the notice on your schedule — nothing is suppressed and no uptime is
touched. Announcement-only windows are one-time only, so a recurring series
can't flood your status page.

## Manage a running window

You don't have to wait for the clock. From the window's **⋯** menu:

- **Start now** — begin an upcoming window immediately (asks you to
  confirm).
- **Extend** — running long? Push the end out by **+15m / +30m / +1h**; the
  suppression and the status-page notice move with it.
- **Complete now** — finished early? End it now — alerts resume immediately
  and the status-page notice is marked completed.

## Times & timezones

Windows are scheduled and shown in **your profile time zone** (avatar →
Settings → Profile), not the browser's: the editor says which zone it is
using, and the list shows the next run in that zone with its **GMT offset**,
the window's own zone (or UTC) on hover and a **copy** action — handy when
your team spans timezones.

## In your Inbox

Every member gets maintenance events in the dashboard bell and
[Inbox](/docs/organization#inbox-and-in-app-notifications): *scheduled* when a
window is created, *started* when it activates (by the clock or by **Start
now**) and *ended* when it ends, is completed early or is disabled. Switch
them off under Settings → Notifications if you don't want them.

## Maintenance vs. pausing vs. muting

| | Checks | Alerts | Uptime / SLA |
| --- | --- | --- | --- |
| **Maintenance window** | keep running | suppressed for a planned period | time excluded |
| **[Mute alerts](/docs/monitors#actions-on-a-monitor)** (1h–7d) | keep running | notifications held | counts normally |
| **[Pause](/docs/monitors#actions-on-a-monitor)** | stop entirely | none | not measured |

> **Info**
>
> Use a **maintenance window** for planned work with a known end time,
> **mute** to quiet a noisy monitor while you investigate, and **pause** for
> a monitor you are stepping away from indefinitely.

## Recurring examples

- **Nightly batch** — Daily, 02:00, 30 min, tag `backend`.
- **Weekly deploy** — Weekly, Sunday 01:00, 1 hour, the monitors behind your
  app.

## Plans

Plans cap how many windows can be **upcoming or active** at once — Free 1,
Starter 3, Pro 10, Business 50, Enterprise unlimited (see
[/pricing](/pricing)). Ended windows are history and never count, so
completing a window frees its slot immediately. The page shows *N of M
windows in use*; at the cap the **New maintenance window** button explains
that you can complete or delete one, or upgrade.

## Also via API

Windows can be listed, created, updated and deleted through the REST API —
see [Maintenance windows API](/docs/api/maintenance-windows). The same
targets (monitor IDs and/or tags), schedule and effect toggles apply.

---

Canonical page: https://watchfor.io/docs/monitors/maintenance · All docs: https://watchfor.io/docs · Site guide: https://watchfor.io/llms.txt