Grafana OnCall open source is finished. Grafana put it in maintenance mode on March 11, 2025 and archived it on March 24, 2026. After that date its APIs and Cloud Connection stopped working with the open source version, so SMS, phone and push relays that depended on Grafana's cloud no longer reach it. Your install keeps running, but nobody patches it beyond critical fixes made before the archive. The realistic replacements are Grafana Cloud IRM, a router such as PagerDuty or Splunk On-Call, or, if your pages are mostly outages, an uptime monitor that pages people itself.
Which of those fits depends on one question: where do your alerts come from? This article walks through the dates Grafana published, what actually breaks on an archived install, and which kind of replacement matches which kind of team, with yearly costs for the options that publish a price.
What Grafana announced about OnCall open source
Everything in this section is from Grafana's own maintenance mode announcement and the GitHub repository README, read in November 2026.
| Date | What happened | What it means for you |
|---|---|---|
| March 11, 2025 | OnCall (OSS) enters maintenance mode | Fixes only for critical bugs and CVEs scored 7.0 or higher, no new features |
| April 21, 2025 | OnCall (Cloud) app sunset in favor of Grafana Cloud IRM | Cloud customers move to the combined OnCall and Incident app |
| March 24, 2026 | OnCall (OSS) archived, Cloud Connection deprecated | APIs and Cloud Connection stop working with OSS, existing installs run on their own |
Grafana states that it is not forcing OSS users onto Grafana Cloud, and the code stays under AGPLv3, so you can fork it. That is a real option for a team with spare platform engineers. For everyone else, an archived paging system is a risk you carry alone: when the SMS relay stops delivering at 3am, nobody upstream will fix it.
What breaks on an archived self-hosted OnCall
Grafana says existing deployments can be used indefinitely, so nothing fails on archive day. The first breakage is the part that was never self-hosted: Cloud Connection supplied the SMS, phone call and mobile push relay, and it is deprecated. Teams that wired their rotation to phone calls through that relay should test a real page this week rather than discover the gap in an incident. The IRM mobile app also worked with self-hosted OSS only until the same date.
The slower risk is security. Maintenance mode already limited fixes to serious vulnerabilities. After the archive there is no maintainer to ship one. A paging tool sits close to your credentials and your integrations, so an unpatched one is worth replacing even if it still pages.
Grafana OnCall alternatives by type of team
There are three honest replacement paths. The right one is decided by where your alerts originate, not by price.
| Replacement path | Best for | Pricing basis | Catch |
|---|---|---|---|
| Grafana Cloud IRM | Teams staying in Grafana, with alerts from Alertmanager and Grafana alert rules | Paid add-on billed on monthly active IRM users | Grafana's docs give no per-user price on the page we read, so get a quote |
| Router: PagerDuty, Splunk On-Call, Datadog On-Call | Teams with alerts from many tools who need one rotation | Per user, about $5 to $21 a month, see the PagerDuty alternative costs | Seats multiply with headcount, and you still need something to detect the outage |
| Uptime monitor with paging: AlertPing | Teams whose pages are outages, failing APIs, expiring certificates, silent cron jobs | Flat, $348 a year on Team for 100 monitors and 15 members | Does not ingest alerts from Prometheus or Grafana, it raises its own |
When an uptime monitor is enough to replace OnCall
A lot of small teams adopted Grafana OnCall because it was free to self-host and they needed something to wake a person up. Look at the last 30 days of pages. If most of them were the site timing out, an API returning 500s, a port closed or a certificate expiring, you were running a heavy paging system to relay messages a simple outside check can produce directly.
AlertPing runs HTTP, API, port, ping, SSL and heartbeat checks every 30 seconds from three regions and confirms a failure from a second region before it pages anyone. Escalation chains page step one, wait the time you set, then move to step two if nobody acknowledges. On-call hours decide who is step one at which time of day. SMS, email, Slack and webhook alerts are on every plan, and no plan charges per user. The detail is on the alerts and escalation page.
What it does not do is take an alert Prometheus fired about CPU or a queue depth. If those are your pages, you need a router, and keeping an uptime monitor next to it is the pattern we describe in uptime monitoring to pair with APM. Teams that also collect customer requests or approvals through forms sometimes run those on a tool such as Formsbuilder, and wire the form endpoint into the same uptime checks so a broken intake form pages someone too.
What replacing OnCall costs per year
Annual-billing list prices multiplied by users and 12 months. Router rates come from the vendors' published pages as logged in our Splunk On-Call pricing and PagerDuty pricing guides.
| Team size | Splunk On-Call | PagerDuty Professional | AlertPing |
|---|---|---|---|
| 5 users | $300 | $1,260 | $348 (Team) |
| 10 users | $600 | $2,520 | $348 (Team) |
| 25 users | Quote | $6,300 | $1,128 (Business) |
The routers do not include detection, so add the monitoring tool you already pay for. The AlertPing column includes both detection and paging, which is why it only fits teams whose pages are outages.
How to migrate off OnCall without missing a page
- Export your schedules and escalation chains from OnCall and list every integration that sends it alerts.
- Sort each integration into "outage" or "internal signal". Outage sources move to the uptime monitor, internal signals need a router.
- Rebuild the escalation chain in the new tool with the same wait times and send a test page to every person on it.
- Run both systems side by side for one week, then switch the old one off.
Start with the monitors: a Team plan lets you paste a URL and a phone number and have a working check in a minute. See the plans and compare with the Splunk On-Call alternative page if a router is on your shortlist.
Is Grafana OnCall still free to use?
The open source code is still free under AGPLv3, and existing installs keep running. What is gone is maintenance, the APIs that connected it to Grafana Cloud and the Cloud Connection relay for SMS, phone and push. Grafana's supported path is Grafana Cloud IRM, a paid add-on billed on monthly active IRM users.
What replaces Grafana OnCall for on-call scheduling?
For full rotation scheduling with overrides across many alert sources, a router such as PagerDuty or Splunk On-Call replaces it. For outage paging with escalation and on-call hours, an uptime monitor with built-in chains such as AlertPing does, at a flat price instead of per seat.