Opsgenie MCP server: archived, and ending

Two independent reasons not to build here, and where Atlassian is pointing Opsgenie customers instead.

Last verified 1 September 2026 · from Atlassian's MCP server documentation, the archived community repository and the official registry

This page is one of 90 in a directory of monitoring and developer tools. Each page checks what an AI coding agent can get out of the tool through its MCP server, the connector that lets the agent query the tool directly.

What is the Opsgenie MCP server, and is it usable?

Opsgenie is Atlassian's alerting and on-call product: it routes alerts, runs escalation policies and manages who is paged.

Atlassian has been consolidating that capability into Jira Service Management, and Opsgenie has a published end-of-life date rather than a roadmap.

That changes the advice. For most vendors in this guide the question is which connector to use. For Opsgenie the question is what to migrate to.

Every one of the 90 vendors in this directory gets the same four checks. Opsgenie passes 0 of them.

Opsgenie works in production observability, and nothing maintained connects to it.

Opsgenie's alerts and schedules are not reachable through anything maintained. The end-of-life date means that will not change. Plan around the destination product rather than this retired surface.

What can an AI agent do with Opsgenie?

Two things to watch, and the second outranks the first.

Do you need the MCP server at all?

No, and no maintained one exists to need. If you are staying with Atlassian, Jira Service Management is where the capability is consolidating, and it is covered by Atlassian's own MCP server alongside Jira and Confluence.

If you are moving off, several alerting and incident platforms publish maintained servers. PagerDuty exposes 101 tools with a read-only default, covering on-call, escalation and incidents. No PagerDuty tool name is listed here. incident.io documents 32 tools including escalation_respond for acknowledging a page and escalation_path_show for who would be paged at each level.

Rootly registers 252 tools covering shift metrics, handoff summaries and responder availability. FireHydrant offers five tools with a read-only scope. Neither set is named tool by tool on this page.

For the alerts themselves, before they reach an on-call tool, Prometheus exposes firing alerts and alerting rules, and Grafana, Datadog and Dynatrace all reach their own alerting surfaces.

Do not build against Opsgenie. Connect the platform you are migrating to, and use its server instead.

Opsgenie MCP server setup

Nothing here installs an Opsgenie server, because there is none and the only community one is archived. This is a migration route, and the date on it matters.

Opsgenie is scheduled for end of life on 5 April 2027. Anything built against it now has a fixed shelf life, which changes the calculation from integrate to replace.

Connect the platform you are moving to, and use its server. Two in this directory publish one.

claude mcp add --transport http incidentio \
  https://mcp.incident.io/mcp

claude mcp add-json "pagerduty" '{"command":"uvx",
  "args":["pagerduty-mcp"],
  "env":{"PAGERDUTY_USER_API_KEY":"<your-key>"}}'

Every other client

Each block below is the configuration for one client, with the file path and the key that client expects.

Atlassian MCP: Jira · Confluence · JSM · Bitbucket · Compass · Rovo
Opsgenie tools: none

gh api repos/giantswarm/mcp-opsgenie --jq .archived   # true

Read-only access, permissions and security

• No Opsgenie credential is worth creating

  With end of life set for 5 April 2027, a new integration credential against Opsgenie has a known expiry.

• Paging is a write action

  A connector that can create or escalate an alert can wake someone up. Scope the key to reads, unless paging someone is the point.

• A fork inherits its own risk

  Forking an archived server means you own its dependencies and its security updates, with no upstream to receive a fix.

• incident.io's hosted route stores no token

  Browser authorization keeps a long-lived credential out of the configuration file.

• On-call data is personal data

  Schedules and escalation policies name individuals and their availability, and that reaches the model with the rest.

Troubleshooting

Atlassian's MCP server is connected but exposes no Opsgenie tools

That is expected. Atlassian's own server covers Jira, Confluence, Jira Service Management, Bitbucket, Compass and Rovo. It carries no Opsgenie tools at all.

You want to fix or extend `giantswarm/mcp-opsgenie`

You cannot upstream it. The repository was archived on 8 May 2026 and accepts no further changes. A fork is the only route, and you would own it.

The MCP registry returns nothing for Opsgenie

There is nothing to return. A registry search on 24 August 2026 found no Opsgenie entry, and no maintained package exists either.

You are deciding whether to build a connector yourself

Weigh it against the end-of-life date. Opsgenie is scheduled for end of life on 5 April 2027, so anything built now has a fixed shelf life.

Migrating the alerting first, then connecting whatever replaces Opsgenie, avoids paying twice.

Opsgenie MCP server: Reference

What engineers report

Two dates carry this section: when the community server stopped, and when the product stops.

Repository facts for giantswarm/mcp-opsgenie come from the GitHub API on 24 August 2026, which reports it archived. Atlassian's MCP server documentation was checked on the same date and lists Jira, Confluence, Jira Service Management, Bitbucket, Compass and Rovo without Opsgenie. The official MCP registry has no Opsgenie entry.

Should you connect an agent to Opsgenie?

No. A server exists and is not usable. Best for migrating, and connecting the destination platform's server instead.

Strongest fit

Nobody. The product has an end-of-life date and no maintained connector.

Main advantage

None to report: the only Opsgenie server is archived.

Main weakness

Two independent blockers, an archived server and a 2027 end of life.

Operational risk

Building an integration with a known expiry date.

Frequently asked questions