H2 Pager · an Opsgenie alternative
The pager,
after Opsgenie.
Opsgenie reaches end of support on 5 April 2027, and Atlassian switches it off that day. H2 Pager does the same job. When something breaks, the phone of the person on call rings, even on silent, and it keeps going until someone answers. Alerts can come from ServiceNow, your monitoring or any other tool. It starts from proven, industry-standard tools and is shaped to fit how your teams work.
Highlighted in green what H2 Pager adds on top of Opsgenie
5 April 2027Opsgenie stops
120 daysOpsgenie keeps running after Atlassian's migration
Ready to onboardfitted to your needs at no extra cost
Agentic AI, no extra costautomation and agent-run workflows on your company's approved models
A preview, with example data
This is H2 Pager's screen, simulated in your browser with made-up teams, people and alerts. It is not connected to anything real. Simulate a P1 and Ioana's phone rings: answer it, let it escalate, or find someone to cover a shift. Time runs ten times faster here.
8 tools → message bus ↔ H2 PagersimulatedExample data, invented peopleEscalations run 10× faster
Fig. 2A preview of H2 Pager, simulated in your browser on example data. It is not connected to any real system.
How it works
Alerts come from many tools. H2 Pager collects them in one place, works out who should get each one, rings that person's phone, and tells the other tools what happened.
Table 1 The path of one alert.
-
01 / sources
An alert arrives
From ServiceNow, a monitoring tool, a security tool or an email. Each tool sends it the way it sends to Opsgenie today.
-
02 / message bus
It is kept safe
A message bus holds every alert until H2 Pager has handled it. Nothing is lost if it restarts, and repeats are dropped.
-
03 / one shape
Every alert looks the same
H2 Pager turns each tool's message into one format: what happened, how urgent it is, which team and which service.
-
04 / routing
H2 Pager finds the right person
The alert names the team. The rota says who is on call right now, including swaps and cover.
-
05 / the phone
The phone rings
An alarm on the Android app, loud even on silent. If nobody answers, a phone call. Then the backup, then the team lead.
-
06 / events out
Everyone is told
Every change (rang, answered, resolved) goes back out through the message bus. ServiceNow updates the ticket, the tool that raised the alert marks it as seen, and a Teams channel can follow along.
| Tool | Connects through | What H2 Pager reads | What it gets back |
|---|---|---|---|
| ServiceNow | A business rule on incident insert and update | Assignment group, priority, configuration item | In Progress, work notes, resolution |
| Azure Monitor | An action group that posts to the topic | Severity, resource, resource group | Alert state Acknowledged |
| Prometheus, Grafana | An Alertmanager receiver or a Grafana contact point | Labels such as team, severity and service | A silence until it resolves |
| Zabbix, Nagios, Icinga | A media type or a notification command | Host, trigger, severity | Problem acknowledged |
| Microsoft Sentinel, a SIEM | An automation rule or a playbook | Incident severity, owner, entities | Incident status and owner |
| A mailbox H2 Pager reads, for backup jobs, appliances and vendors | Sender and subject rules | Nothing; its alerts clear by rule | |
| Anything else | An HTTPS webhook, or a message put straight onto the topic | A mapping written once | A callback, or a subscription of its own |
A watchdog on separate infrastructure. If a tool goes quiet for too long, or H2 Pager itself stops answering, the NOC lead gets a text message from a separate service. A pager has to work on the night everything else is failing.
Who gets paged, and when
Urgent alerts ring at any hour. Less urgent ones wait for office hours or stay on the screen. These are the starting rules; each team can change its own.
| Priority | Pages | Otherwise |
|---|---|---|
| P1, P2 | At any hour: an alarm, then a call | |
| P3 | Office hours, Monday to Friday 09:00 to 17:00 | Waits on the dashboard; pages the next shift if still open |
| P4, P5 | Never | Dashboard only |
| After | Who is reached | How |
|---|---|---|
| 0 min | The person on call | Alarm on the phone |
| 2 min | The person on call | Phone call |
| 10 min | The backup | Alarm and call |
| 20 min | The team lead | Alarm and call |
Most operations teams work one of two ways, and H2 Pager handles both. NOC and SOC work 12-hour shifts around the clock. The other teams have someone on duty in office hours and one person on call at night and at weekends, changing every week, with a backup. Swaps and cover are handled in H2 Pager. Holidays are on the calendar too, so a clash shows up weeks before anyone misses a call, and cover is offered first to whoever has done the fewest nights. The calendar reads like Opsgenie's: the day and night rotations, the overrides on top, the backup, and the final schedule that results.
When something breaks
A pager matters most on the worst night. Each of these has a plan, and the dashboard shows the state of every part.
The phone doesn't get the alert. A phone call follows after two minutes. Every phone gets a test alert each week, so a broken setup is found early.
Nobody answers. H2 Pager calls the backup, then the team lead. Every step is recorded with its time.
A tool goes quiet. Every tool sends a regular sign of life. If one stops for too long, the NOC lead gets a text message.
An alert can't be read. It is set aside, never thrown away, and shown on the dashboard to be fixed.
Updating another tool fails. If ServiceNow is down, the update waits and is tried again. Ringing the phone never waits for it.
A cloud region goes down. H2 Pager runs in two regions. The watchdog and the phone-call service run on separate infrastructure, so the alert about the outage still gets through.
What you keep, and what you gain
H2 Pager does what teams use Opsgenie for today. The rows marked New are what Opsgenie does not do.
| What it does | Opsgenie | H2 Pager |
|---|---|---|
| Carried over | ||
| On-call rotas, with overrides and swaps | Yes | Yes |
| Escalation to the backup, then the team lead | Yes | Yes |
| An alarm on the phone, then a phone call | Yes | Yes |
| Alerts from ServiceNow, monitoring and security tools | Yes | Yes, through the message bus |
| Acknowledgements and updates back to those tools | Yes | Yes |
| A warning when a tool goes quiet | Yes | Yes |
| Reports on alerts and response times | Yes | Yes, the dashboard |
| The phone app | Android and iPhone | Android; iPhone later |
| New On top of Opsgenie | ||
| Holidays on the calendar, with clashes flagged weeks ahead | No | Yes |
| Cover offered fairly: each person's night shifts this month are shown | No | Yes |
| Every phone tested every week, automatically | No | Yes |
| Adapted to your teams' specific needs, at no extra cost | No | Yes |
| Agentic AI workflows on your company's own approved models, at no extra cost | No | Yes |
| Still paging after 5 April 2027 | No, switched off | Yes |
| New After the move (section 07) | ||
| The changes planned during your shift, and the change behind an alert | No | After the move |
| A handover at every shift change | No | After the move |
| New tasks and requests sent to whoever is on duty | No | After the move |
Migrating from Opsgenie
Atlassian’s own migration leads to Jira Service Management or Compass. H2 Pager is the other way off Opsgenie, and nobody relies on it until it has proved itself on your own alerts. Opsgenie keeps paging exactly as it does today while H2 Pager runs alongside it, and a report shows where the two agree.
Table 6 The move, step by step.
-
01 / week 1
H2 Pager reads your Opsgenie
With a read-only key, H2 Pager copies your people, teams, rotas, overrides and escalation rules. Nothing in Opsgenie changes, and no phone does anything new.
-
02 / weeks 1 to 2
The calendars are checked
Opsgenie's own on-call timeline and H2 Pager's are compared hour by hour, eight weeks ahead, every day. People keep editing rotas in Opsgenie, and H2 Pager picks up each change within 15 minutes.
-
03 / week 2
H2 Pager gets a copy of every alert
H2 Pager is set up alongside Opsgenie. Your tools keep sending to Opsgenie exactly as today, and a second rule in each one sends it the same alert through the message bus.
-
04 / weeks 3 to 5
H2 Pager listens and rings nobody
For each alert, H2 Pager works out who it would page and when, and changes nothing in ServiceNow. A daily report compares that with what Opsgenie did. Meanwhile the screens, the dashboard, the updates back to ServiceNow and the watchdog are finished and tested.
-
05 / weeks 6 to 7
A few volunteers carry both
A handful of people get H2 Pager's alarm as well as Opsgenie's, marked as a test. Real alerts, real phones and real calls, while Opsgenie still does the paging that counts.
-
06 / weeks 8 to 10
Teams move one at a time
When a team's report is clean, H2 Pager takes over its paging and starts updating ServiceNow. Opsgenie stays as the fallback until the last team has moved, and is then switched off, well before 5 April 2027.
| Check | What is compared | How often | Passed when |
|---|---|---|---|
| Setup | Every person and contact method, team, rota, override, escalation and routing rule, personal notification rule, and the ServiceNow mapping | In week 1, then after every change | Each one is carried over, or the difference is agreed |
| Calendars | Who is on call and who is backup, hour by hour, eight weeks ahead, for every team | Every day | Every hour names the same people |
| Alerts | For each alert: received, grouped with its repeats, team, priority, first person paged, each escalation step, time to the phone | Every alert, from week 3 | Every alert matches or the difference is agreed, and H2 Pager is no slower |
Three kinds of difference. The same; different by design, because you chose it (letting P3 alerts wait for office hours, for example); or wrong, which is fixed before that team moves. Nothing counts as proof until it is sorted.
The report stays with you. It holds real names, phone numbers and alert text, so it stays private to you. You keep it after the move, as the record of how the switch was tested.
After the move: the work around the pager
New H2 Pager knows who is on shift, who is on leave and who has the daytime duty. Once the move is done, it can use that to show the work that matters during a shift. This comes after the move, not as part of it.
Changes during your shift. The on-call screen lists the changes planned while you are on call. When an alert comes in, a change made to the same system shortly before is shown with it, because that is often the cause.
A handover at every shift change. At 08:00 and 20:00 the next person gets a summary: open incidents, changes in progress, tasks waiting on the team, and a short note from the person going off shift.
New work to the person on duty. A new task or request goes to whoever has the duty for that team right now, not to the whole group. Taking it, or adding a work note, writes straight into ServiceNow.
Three rules for the work around the pager.
- ServiceNow stays the record.H2 Pager reads tasks, requests, changes and the systems they touch, and writes notes back into the ticket. It keeps no copy of the work and holds no asset inventory of its own.
- Work never rings.Tasks live in a separate, quiet part of the app, with their own low-priority notifications on Android, so no phone setting can make a task ring like an alarm. Some people prefer a summary a few times a day instead.
- After the move, not before.The move away from Opsgenie stays small and proven. Reading tasks and changes needs more access in ServiceNow, so this gets its own review.
Built on the industry standard
(i) We start from what the industry already trusts.
A message bus carries the alerts, a well-tried on-call engine keeps the rotas and the escalations, a push service delivers the alarm and a voice service makes the calls. Nothing exotic, and nothing your teams have to learn from scratch. We set it up, look after it and are there when something goes wrong.
The parts
Names, rotas and alert history are yours, and you can take them with you.
Made to fit your teams
From that standard base we build the version that fits how your teams work, often at a lower cost. Adapting it to your needs costs nothing extra, and neither do automation and agentic AI workflows on the models your company already trusts. H2 Pager measures itself and points out what could work better, so better results are something you can see, not a promise.
Table 8 Three ways H2 Pager keeps getting better for you.
-
01 / your requests
What your teams ask for, built
Your teams tell us what would help: a new source, a screen, a rule, a report. We build it, at no extra cost. What helps everyone goes into H2 Pager itself, and what is particular to you goes into your setup.
-
02 / analysis
H2 Pager looks for what could be better
It reads its own history and points out what to improve: alerts nobody ever needs, alerts that wake someone and are closed within minutes, a system that fails the same way every week, nights that fall on the same few people, hours when answers come slowly. Each finding comes with a suggested change, and nothing changes until your team agrees.
-
03 / ai agents
Automation and agentic AI workflows
Automated workflows, and agentic ones where your AI agents carry out the steps within the rules your team sets, come at no extra cost. They run on the AI models and agents your company has already vetted and approved. An agent can gather what is known when an alert arrives, draft the work note and the handover, or run the checks your runbook allows.
How agents work in H2 Pager.
- Your models and agents, not ours.Only models and agents your company has approved. We connect them, they run under your company's own AI agreement, and nothing goes to an AI service you have not chosen.
- The same doors as people.An agent works through the same interface as a person, with its own name and its own permissions. Every step it takes is recorded as an event, next to everyone else's.
- People decide.An agent can gather, draft and suggest. Paging someone, closing an incident or changing a rota happens only under rules your team has set.
Switch before 5 April 2027
A ten-week migration off Opsgenie, from a read-only look at your setup to Opsgenie switched off, well before Atlassian deletes whatever is left in it on 5 April 2027. Nothing changes for your people until the numbers say so.
- Weeks 1 to 2H2 Pager reads your Opsgenie setup and checks every rota against it, hour by hour. A real call to a real phone.
- Weeks 3 to 5H2 Pager listens to every alert and rings nobody. A daily report compares it with Opsgenie.
- Weeks 6 to 10A few volunteers carry both, then teams move one at a time. Opsgenie stays as the fallback.
- ThenOpsgenie switched off, well before 5 April 2027. You keep the setup guide, the reports and the people who know it.