An alert fires — incident or planned work?
Open it: Grafana upgrade is running. Ping the owner, not the on-call.
Maintenance window management for engineering teams
See every maintenance — running, done, scheduled — without the Slack-and-Jira hunt.
MaintMode is maintenance window management for DevOps, SRE and infrastructure teams — the services, databases and clusters you run, not the building you sit in. Plan maintenance windows, coordinate resources, detect conflicts, and keep every operational change on one shared timeline your team actually keeps current.
A 30-second form — no account needed · live instance · open source, free to self-host.
Built by engineers who've been paged at 3 a.m. one too many times.
The problem
Maintenance work is scattered everywhere.
Buried in Slack
Someone maybe posted in #maintenance. Maybe in a DM. Good luck finding it later.
Lost in Jira
Change tickets exist, but no one scans them to see what's happening right now.
Or nowhere at all
Half the work just happens. Nobody has the full picture of what's going on.
See everything · at a glance
Open it — and the whole picture is right there.
Past
What just finished
Running
What's happening now
Scheduled
What's coming
- Plan → Approve → Execute
- Resource-aware conflict detection
- Full audit trail
- Roles & approvals
Where it helps
A few moments it pays off.
Scheduling a database migration tonight
It flags a cluster drain already booked at the same hour — before you commit.
See it screen by screen
A 6-step walkthrough of one maintenance window, from setup to approval — with real screenshots.
Maintenance window conflict detection
It spots the conflict so you don't go hunting.
MaintMode detects conflicts between maintenance windows that overlap in both time and affected resources, and reports them before the work is approved. No checking tickets, no asking around. A shared calendar can't do that — it knows about time, but not about the database two teams just booked for the same hour. See it on a real screen.
The honest part
A timeline is only useful if it's actually current.
Most tracking dies because nobody keeps it up to date. MaintMode is built to be the one place worth updating — fast to log a window, conflict checks that reward entering work early, and a view the whole team relies on. The more it's used, the more accurate it gets.
Open core
Run it yourself, or let us run it for you.
MaintMode is open core: the full product is free to self-host, and the hosted option runs one isolated VM per organisation rather than a shared multi-tenant database. Same product either way — the hosted one just means you don't operate it.
Self-hosted
Free, on your own infrastructure
The full product — not a crippled community edition. Your data never leaves your servers. Source is on GitHub under AGPL-3.0.
Hosted
One isolated VM per organisation
Not a shared database with a tenant column — your organisation gets its own machine with its own Postgres. Same product as the self-hosted one, we just operate it.
Why not just…
You've already tried the alternatives.
A shared calendar
Shows events, not resources. No conflicts, no ownership, no audit trail.
Slack + Jira
The truth is scattered. You reconstruct it by hand, under pressure.
Jira Service Management
Has a change calendar with maintenance windows and conflict warnings — on Premium, once someone configures it. Built around change requests and approval queues; MaintMode is built around the resource a window touches.
ServiceNow & co.
Heavy, expensive, needs an admin. Overkill for one job.
MaintMode
One timeline. Resource-aware. Conflict-aware. Light enough to actually use.
Questions we get
Straight answers, including the noes.
Is this an incident management tool?
No. MaintMode is for planned work — maintenance windows you schedule, review and run. It has no on-call rotations, no paging, no incident timelines. If something is on fire, this is the tool that tells you whether the fire was scheduled.
Does it integrate with our alerting or monitoring?
Not today. MaintMode doesn't ingest alerts and doesn't watch your systems. Windows are entered by people. The value is that when an alert does fire, there's one place that answers "is anyone working on this right now?"
How is this different from a shared calendar?
A calendar knows about time. MaintMode also knows about resources — the specific services, databases and clusters a window touches — so it can detect that two windows target the same one at the same hour and say so before either is approved.
Can we self-host it?
Yes, and it's free. The core is open source under AGPL-3.0 — backend, UI and the self-hosting setup are on GitHub. The hosted option gives each organisation its own isolated VM rather than a shared multi-tenant database.
Does it do change freezes or blackout windows?
No. MaintMode doesn't implement change freezes today. It detects conflicts between scheduled windows; it won't stop you from booking work during a company-wide freeze period, because it has no concept of one.
What does it cost?
Self-hosting will be free. Hosted pricing isn't set yet — we're in early access and working with the first teams to figure out what makes sense. Nothing to pay right now.
Stop guessing what's being worked on.
There's a live instance. Tell us a bit about your team and we'll get you in.
Request accessPrefer to run it yourself? The core is going open source — free to self-host.