Maintenance windows that keep the night quiet.

Schedule a window and Uptimia pauses the checks on the monitors the work touches, holds every alert, and keeps the planned span out of your uptime figures. One-off or recurring, announced on your status page, and it closes itself when the work is done.

Start free trial → Browse plans
30-day free trial · one-off & recurring windows No credit card Schedule a window in under a minute
Monitor types
12
Recurrence patterns
4
Checkpoints
171+
Countries covered
70+

The migration nobody was paged for

02:00 on a Sunday, Caldmont took its database down to migrate it. At 02:15 the database went down and the storefront answered 500 — the checks were paused, the on-call phone stayed dark.

02:00 The window opens — checks pause on the two covered monitorsScoped to Storefront and Orders API — nothing else on the fleet changes in window
02:15 The migration takes the database downAny other night this pages the on-call — the window suppresses it in window
04:00 The window closes — checks resume on their ownMonitors go green on the next tick — no switch to remember to flip back in window
04:01 The two hours never touch your uptime figureExcluded from the SLA denominator — not counted as downtime, not inflated as uptime excluded
0pages, all night
Planned work — and nobody was paged.04:00
Skip the window and a 2 a.m. deploy pages the on-call for downtime you caused on purpose — then leaves a two-hour scar in the report you send customers. The window paused the checks, held every alert, and kept the planned span out of your numbers.
checks pausedalerts heldexcluded from uptimeone-off or recurring · scoped to the monitors you pick
And without a window? The deploy you scheduled pages the on-call at 2 a.m. — for downtime you planned. The team learns that pages can be ignored, and the availability report shows an outage that never really mattered. a false scar

Scope it, then forget it

One window against a migration, a nightly job or a whole group of monitors. It opens on its schedule, holds the alerts, and closes without anyone remembering to flip a switch back.

One-off windowsA single deploy or migration
Recurring windowsDaily, weekly or monthly
Scoped to the monitors you pickAll, by type, by group or by hand
Alerts suppressedNo pages, no escalation ladder
1 1 window against any monitor
Excluded from uptimeKept out of the SLA figure
Shown on your status pageCustomers see it was planned
Ends early when you finishEnd maintenance from the dashboard
The schedule is computed in your account time zone: a window that crosses midnight, lands on the last day of an uneven month, or spans a daylight-saving change runs exactly when you meant it to — not an hour early or a day late. planned, not paged
Scheduled

Schedule the window before you deploy

Pick the monitors a change will touch, choose one-off or a recurring pattern, and set the start and end. A live preview shows the next occurrences before you save.

  • One-off or recurring — a single window, or a daily, weekly or monthly pattern
  • Scope it four ways — all monitors, by type, by group, or hand-picked
  • A live preview shows the next occurrences and warns of any overlap before you save
See how maintenance windows work →
New maintenance windowrecurring · weekly
1patternEvery Sunday✓
2window02:00 – 04:00✓
3scopeStorefront + Orders APIediting
4announcePost to status page✓
5previewNext: Sun 02:00✓
＋ add monitors — all, by type or group
Previewlive
Next: Sun 02:00 – 04:00
on schedule
The preview resolves the next occurrences in your account time zone — the same engine that runs the window, so what you see is exactly what gets suppressed.
next 5 occurrencesoverlap checked
Reschedule and the public thread moves with it — edit a published window's time and the status-page announcement follows, no stale slot left advertised. preview

Alerts held, SLA clean, customers told

Alerts stay quiet

The on-call sleeps through your deploy

While the window is open, checks on the covered monitors are paused and every alert is held — a planned deploy doesn't wake anyone, and doesn't start an escalation ladder.

  • Checks paused for the duration — probes skip the covered monitors
  • Alerts suppressed — no pages, and a queued down-alert that has not gone out yet is cancelled
  • Open incidents force-closed at the start, so a flapping deploy can't spam
How alert suppression works →
Window · Storefront · caldmont.com02:00 · running
02:00opened
02:15check skipped
02:15alert held
02:50check skipped
04:00closed
What the check saw at 02:15
500
Down — but in maintenance
No alert sent
Storefront answered 500 at 02:15 during the migration. The check saw it; the on-call didn't — the window held the page.
check saw itno alert sent
Neither did the escalation ladder. A suppressed incident never starts a policy, so nobody up the tiers is paged either. held
Reporting

Planned downtime stays out of your SLA

A maintenance window isn't 100% up and it isn't an outage — it's excluded. Uptimia subtracts the window from the availability denominator, so planned work neither dents your uptime nor inflates it.

  • Excluded from the denominator — availability is measured over non-maintenance time
  • On private reports and the public status page alike — the same rule in both places
  • No double-counting — a suppressed incident already contributes zero downtime
How uptime is calculated →
"We were down two hours"
no — that was planned, and excluded
Maintenance excluded — this month30 d
The month's windows, subtracted from the uptime denominator — the migration, the nightly job, the DNS change.
DB migration2 h
Cache flush45 m
TLS rotation20 m
DNS change10 m
Config push5 m
excluded from SLAprivate + publicper window
Availability reads 99.98%, not 96.7%. The two-hour migration is excluded — the figure reflects unplanned downtime only. excluded
At a glance

Every occurrence on one timeline

One timeline answers the questions that matter before a deploy: what is running now, what is coming up, and which monitors each window covers.

  • Running now, up next and recurring up top — every window in one place
  • Which monitors each window covers, and the pattern in plain language
  • A 7- or 14-day timeline with each occurrence positioned by real time
Tour the maintenance page →
?What is running now?
?What is coming up?
?Which monitors are covered?
Maintenance — caldmont.comlive
Running now
1 window
Storefront + Orders API
Ends in
1:19
Sun 04:00
Next window
Sun 02:00
weekly · migration
Timeline
14 days
every occurrence
The timeline is the engine — every block is positioned by the same occurrences that pause your checks, exact across DST and month lengths. 7 / 14 days
Status page

Customers see planned, not broken

Announce a window and it appears on your public status page ahead of time — so when the covered services go quiet, visitors see scheduled maintenance, not an outage.

  • Posted ahead of time — the window shows on the page before it starts
  • Recurring windows roll forward — the next occurrence is announced automatically
  • Ends when monitoring resumes — early or on time, the banner clears
See public status pages →
Scheduled maintenanceSun 02:00
caldmont.com status page — "Database migration, Sun 02:00–04:00." Posted Friday; the banner clears when checks resume.
posted aheadauto-clears
Public pageon your domain
Publish toauto, or pick pages
Subscribersemailed once, ahead
Announcementshown in advance
Announced Friday, cleared 04:00. Visitors saw a planned window, not two hours of red on the status page. planned

Where the alert goes once the window closes

Nothing pages while a window is open. The moment it ends, a check that has not recovered reaches every channel your team already watches.

On-call & escalation
PagerDuty Webhooks Twilio
Direct
Email SMS

12 channels, one contact list — the window decides when you hear, not where.

Browse the full integrations directory →
02:05 · window closed — api.caldmont.com still failing
#ops-alertsSlack
⚠ Still failing — api.caldmont.com · window ended 02:00
held 90 minnobody paged during the windowAcknowledge ↩
+371 ··· 4082SMS
Uptimia: STILL FAILING api.caldmont.com. The maintenance window ended 02:00 and the check has not recovered.
InboxEmail
⚠ Still failing after maintenance — api.caldmont.com
Alerts were held from 00:30 to 02:00. The check was still failing when the window closed, so the incident opened at 02:05…
ProductionPagerDuty
TRIGGEREDStill failing after maintenance — api.caldmont.com
assigned to on-call · via Uptimia integration

The next 2 a.m. deploy shouldn't page anyone.Because you planned it.

Maintenance windows come with every plan, the trial included — one-off and recurring, across all 12 monitor types, announced on your public status page.

Try Uptimia free for 30 days →
30 days free no credit card cancel anytime

How maintenance windows work

Schedule a window against the monitors a change touches — one-off or recurring — and Uptimia handles the rest.

Step 1schedule

Pick the monitors, set the time

Choose the monitors a change touches, pick one-off or a recurring pattern, and set the start and end in your account time zone.

OnceWeeklyMonthly
Monitors
Storefront + Orders API
When
Every Sun · 02:00 – 04:00
Step 2it runs

Checks pause while it is open

When the window opens, checks on the covered monitors pause and every alert is held. A recurring window repeats on its pattern; a one-off runs once and expires.

Status
Running · checks paused · alerts held
Repeats
every Sunday
CancelSave window →
Step 3automatic

Monitoring resumes on its own

When the window ends, checks resume automatically and the planned span is excluded from your uptime figures. Finished early? End maintenance for its monitors from the dashboard and checks come back at once.

#caldmont.com/status
Uptimia 04:00
✓ Maintenance complete — Storefront · caldmont.com
checks resumedexcluded from SLA
Also sent to EmailPagerDuty

Recurrence, bulk actions and exclusions

Recurring patterns, exact

Daily, weekly — even the Nth weekday of the month — or monthly by day-of-month. Computed in your account time zone and exact across leap years, month-ends and daylight-saving changes.

DailyWeeklyMonthly

End a window early

Finished ahead of time? End maintenance for the window's monitors from the dashboard — checks resume at once, and the public thread closes with them.

ended early · checks resumed ✓

Bulk from the dashboard

Select several monitors on the dashboard and schedule one window that covers all of them.

select · maintain · 6 monitors

Excluded from SLA

Maintenance seconds are subtracted from the uptime denominator, on private reports and public pages alike.

denominator · maintenance excluded ✓

Escalation-aware

A suppressed incident never starts an escalation ladder — planned work never climbs the tiers.

in maintenance · ladder never starts

Windows cover every monitor type

From uptime, SSL and speed to DNS, heartbeat and server — one window covers any mix of all 12 monitor types, and respects the groups a teammate is scoped to.

Storefront · caldmont.comUPTIME api.caldmont.comSSL Nightly backupHEARTBEAT

What is a maintenance window?

A maintenance window is a scheduled period during which a monitor's checks are paused and its alerts are suppressed, so planned work — a deploy, a migration, a backup — doesn't trigger a false incident. The window is excluded from your uptime figures and can be shown on your public status page, so planned downtime is recorded as planned, not as an outage.

While the window is open

How does a maintenance window work?

Uptimia
maintenance window
pauses checks · holds alerts
Your monitors
paused, not paged

While the window is open, checks on the covered monitors are skipped and any alert is held — a planned deploy never wakes the on-call.

When work overruns

Still no false page

02:15
deploy 500s
still suppressed
Uptimia
no incident opens

a check fails at 02:15 during the migration → the window suppresses the alert and the span is excluded from uptime

The difference

Pausing a monitor vs. a maintenance window

You could pause the monitor by hand — but you would forget to un-pause it, and the downtime would vanish from your history entirely. A window ends itself, keeps the record of what happened, and takes the planned span out of the denominator instead of out of your history.

See alerting & notifications →
What happensPause by handMaintenance window
Alerts stay quiet during the work✓ Yes✓ Yes
Resumes on its own✗ Manual✓ Automatic
Planned span excluded from uptime✗ Just gone✓ Excluded, recorded
Shows customers it was planned✗ Nothing✓ On the status page
Repeats for a nightly job✗ Every time✓ Recurring

Maintenance windows FAQ

01What is a maintenance window?+
A scheduled period during which a monitor's checks are paused and its alerts are suppressed, so planned work — a deploy, a migration, a backup — doesn't page your team or open a false incident. You schedule it against the monitors the work affects, one-off or recurring, and the window is excluded from your uptime figures.
02Does maintenance affect my uptime percentage?+
Maintenance time is excluded from the uptime denominator — not counted as downtime, and not inflated as 100% up. Availability is measured over non-maintenance time, so a planned two-hour migration doesn't dent your SLA. The same rule applies on private reports and on the public status page.
03Can windows be one-off or recurring?+
Both. A one-off window runs once for a specific date and time. A recurring window repeats daily, weekly (optionally on specific weeks of the month — the Nth weekday), or monthly on chosen days-of-month. The schedule is computed in your account time zone and is exact across leap years, month-ends and daylight-saving changes.
04Does a window suppress alerts and escalations?+
Yes. While the window is open, no alerts go out for the covered monitors — a queued down-alert that has not yet been sent is cancelled — and a suppressed incident never starts an escalation ladder. Open incidents on those monitors are force-closed when the window begins, so a flapping deploy can't spam your team. If a monitor was already down and alerted before the window opened, the recovery it is owed is still sent once a check finds it up again.
05Will customers see it on my status page?+
If you choose to announce it, yes. An announced window is posted to your public status page ahead of time, so visitors see scheduled maintenance rather than an outage — on pages where the scheduled-maintenance section is switched on, which it is by default. Recurring windows roll the announcement forward to the next occurrence, and the banner clears when monitoring resumes, whether the window ends on time or early.
06Can I end a window early?+
Yes. If you finish ahead of schedule, end maintenance for the window's monitors from the dashboard — singly or in bulk — and their checks resume at once. The public status thread is updated with them, so the page stops showing maintenance the moment monitoring is back. For a recurring window, ending early affects only the current occurrence; the next one runs as normal.
07Which monitors can a window cover?+
Scope it four ways: all monitors, by type, by group, or hand-picked. A window can cover all 12 monitor types — uptime, SSL, speed, transaction, DNS, API, domain, virus, blacklist, real-user, heartbeat and server. There is no by-type flag for server monitors specifically, but you can cover them via All monitors, an individual pick, or a group that contains them.
08Does it handle time zones, DST and month-ends?+
Yes. Windows are computed in your account time zone. A window that crosses midnight is anchored to its start day; a monthly window set for a day past the month's end runs on the last day (Feb 29 in leap years, otherwise the 28th); and a window spanning a daylight-saving change runs the right wall-clock hours. The schedule you see is the same one that suppresses your checks.
09Does maintenance delete my downtime history?+
No. Maintenance is excluded from the uptime denominator, not erased — a suppressed incident simply contributes zero downtime. Deleting a window reverts its monitors, and resolved public maintenance announcements are preserved so your status-page history stays intact.
10Maintenance window vs. pausing a monitor?+
Pausing a monitor stops it indefinitely and by hand — easy to forget to undo, and the downtime just disappears from your history. A maintenance window is scheduled and self-ending: it silences alerts only for the planned period, resumes monitoring on its own, records the span, and subtracts it from the uptime denominator rather than deleting it.
11Can I schedule maintenance in bulk?+
Yes. You can put several monitors into maintenance in one action straight from the dashboard, and a single window can cover a whole group or type at once. Windows sit alongside every other monitor setting, sharing your contacts, groups and roles.
12Which plans include maintenance windows?+
All of them. Maintenance windows come with every Uptimia plan — the 30-day trial and the free plan too — with the same recurrence engine, SLA exclusion and status-page announcements everywhere. There is no per-window charge and no per-seat billing: a window that quiets a whole team's pagers costs nothing beyond the plan itself.

Schedule your next deploy.

A window takes under a minute to set up, one-off or recurring, and stops your team being paged for downtime you planned.

Schedule a window →
30-day free trial One-off & recurring windows No credit card Excluded from your SLA
Maintenance windows come with every Uptimia plan and cover all 12 monitor types.