--- title: "Atlassian Opsgenie Integration" description: "Atlassian Opsgenie is an incident alerting and on-call platform. Agents create, acknowledge, close, and escalate alerts, open incidents with timelines and postmortems, look up who is on call, and manage schedules, teams, and escalations." url: https://flowrunner.ai/integrations/opsgenie date_modified: 2026-09-04T14:59:52-07:00 --- # Atlassian Opsgenie [Developer Tools](https://flowrunner.ai/integrations/category/developer-infrastructure) Atlassian Opsgenie is an incident alerting and on-call platform. Agents create, acknowledge, close, and escalate alerts, open incidents with timelines and postmortems, look up who is on call, and manage schedules, teams, and escalations. [Verified](https://flowrunner.ai/integrations/verified "What does verified mean?") · 156 actions · API key · available [Atlassian Opsgenie website](https://www.atlassian.com/software/opsgenie) · [Platform Documentation](https://docs.opsgenie.com/docs/api-overview) · Capability data verified 2026-08-25 1. An approved leave request arrives from the HR system 2. Agent resolves the person and reads List User Schedules for every rota they appear in 3. Get Schedule Timeline expands those rotations into the actual shifts across the leave dates, overrides applied 4. The shifts that would be left with nobody on them are identified 5. List User Contacts confirms each candidate for cover can actually be paged 6. The proposed cover, named shift by shift, posts to the team channel 7. The team lead approves before Create Schedule Override changes who gets woken up ## What This Integration Enables Opsgenie is the machinery that decides who gets woken up. Alerts arrive from monitoring, policies rewrite and route them, escalation chains decide who is told next when nobody answers, and rotations decide who that is. Every part of it exists to reach a human at the moment a machine has run out of judgment, which makes it an unusually direct fit for [Orchestration as a Service](https://flowrunner.ai/concepts/orchestration-as-a-service). FlowRunner and Opsgenie are arguing the same thing from different ends: automation is worth having right up to the point where somebody has to decide, and the interesting engineering is in that handoff. FlowRunner [connectors](https://flowrunner.ai/concepts/connectors) cover the whole surface. Agents raise and de-duplicate alerts, acknowledge and escalate them, declare incidents and keep their timelines, open postmortems, read who is on call right now and who takes over next, and manage the schedules, rotations, overrides, teams, escalations, heartbeats, services, maintenance windows, integrations, and policies underneath. Most of that is safe to run unattended, and a great deal of it is better run unattended than by a person at 3am. The exception is anything that changes who can be reached, because that is the one class of change in this platform whose failure mode is silence. ### Without FlowRunner **Rota changes made by hand**: Leave, swaps, and departures are applied one schedule at a time, if anyone remembers **Coverage gaps discovered at 3am**: A rotation with nobody in it looks exactly like a rotation that is fine, until an alert reaches nobody **Alerts arrive without their context**: Whoever is paged starts by gathering the facts the monitoring system already had ### With FlowRunner **Rota changes proposed with their consequences**: Agents expand the real roster and name the hours a change would leave empty **Coverage checked before it is needed**: Contacts, rotations, and escalation membership are verified as part of the change, not after an incident **Alerts arrive investigated**: Details, notes, and the runbook the agent already followed are on the alert before a phone rings ## Use Case Scenarios ### The alert that arrives already investigated A monitor in [Datadog](https://flowrunner.ai/integrations/datadog), [Grafana](https://flowrunner.ai/integrations/grafana), or [Splunk](https://flowrunner.ai/integrations/splunk) fires. Before anything reaches a phone, the agent runs the first three steps of the runbook: it queries the affected service, checks the recent deployments in [Jenkins](https://flowrunner.ai/integrations/jenkins) or [GitHub](https://flowrunner.ai/integrations/github), and looks for a matching error signature. Then it calls Create Alert with an alias, which is the de-duplication key and the difference between a monitor that reports a problem and one that pages somebody forty times. It uses Add Alert Details to write the failing host, the metric value, and the deploy it correlates with onto the alert itself, and Add Note To Alert to record what it checked and ruled out. The person who answers reads a finding, not a symptom. ### Deployment windows that close themselves A release workflow calls Create Maintenance before it starts, so a planned restart does not page whoever is on call. The important half is the ending. Because a maintenance window suppresses genuine failures as well as expected noise, the workflow calls Cancel Maintenance on its success path and on its error path, and Change Maintenance End Date when a deploy runs long rather than stacking a second window on the first. A failed deploy that leaves alerting suppressed is the worst outcome available, so the agent treats closing the window as part of the deploy rather than as cleanup. ### The incident record that writes itself as it happens When an alert crosses a threshold, the agent calls Create Incident against the affected service, which is what starts impact tracking and puts it on the status page. Through the response, every automated action the workflow takes gets a Create Incident Timeline Entry stamped with when the thing happened rather than when it was recorded. Updates post to [Slack](https://flowrunner.ai/integrations/slack) and [Statuspage](https://flowrunner.ai/integrations/statuspage). When impact stops the agent calls Resolve Incident, which is the state a status page reads, then Get Incident Associated Alerts to close the alerts the incident rolled up, and Create Postmortem while the timeline is still fresh. A review that exists is worth more than a review that was going to be written on Monday. ## Human-in-Loop Highlight Update Schedule Rotation replaces the entire participant list in the order it is given. Not merges, replaces. That is how somebody is quietly dropped from an on-call rotation, and nothing in Opsgenie reports it, because a rotation with a hole in it looks identical to a rotation that is fine right up until an alert reaches nobody at 3am. Remove Team Member and Delete User have the same property: the rota does not update itself. So no FlowRunner workflow writes a rotation without showing its consequence first. The agent calls Get Schedule Timeline, which is the only operation that expands rotations into the shifts people will actually work with overrides applied, and posts the difference: "Removing Priya from the Payments rota to cover her leave leaves Saturday 00:00 to 08:00 with nobody on call, and Dan is the only remaining participant with a verified contact method. Add Dan as an override for that window, or extend Sam's shift?" The lead answers, and only then does Create Schedule Override run. The agent did the expansion, the gap detection, and the pageability check. A person decided who loses their weekend. Agent processes routinely Detects exception requiring judgment Clear match Continues automatically Ambiguous Routes to human via preferred channel Human decides Agent resumes with decision ## Agent Capabilities 156 actions ### Alerts 29 - **Create Alert** Raises an alert, which is what notifies whoever is on call. Creation is asynchronous and returns a request id rather than the alert. Setting an alias makes it the de-duplication key, so a repeating condition updates one alert instead of opening dozens. - **Get Alert Request Status** Turns the request id from an asynchronous alert write into its outcome, including the alert id a creation produced. The missing half of every alert write. - **Get Alert** Returns one alert in full. The identifier type has to match what you are passing, because the same path accepts a UUID, an alias, or a short number. - **List Alerts** Returns alerts filtered with Opsgenie's own query language. Without a query this is the whole account newest first. - **Count Alerts** Returns how many alerts match a query without returning them. The right call for a health check or an escalation guard. - **Acknowledge Alert** Marks an alert acknowledged, which stops its escalation. It does not close the alert. - **Unacknowledge Alert** Returns an alert to unacknowledged, which resumes escalation. The correction when an automated acknowledgement turned out to be wrong. - **Close Alert** Closes an alert. Closing resets de-duplication, so the next alert with the same alias opens a new record rather than updating this one. - **Snooze Alert** Suppresses an alert's notifications until a given time. It silences rather than resolves, and escalation resumes when the snooze ends. - **Escalate Alert** Pushes an alert to an escalation policy immediately rather than waiting for the policy's own timer. - **Assign Alert** Gives an alert an owner. Assigning is not acknowledging, so the alert keeps escalating until somebody acknowledges it. - **Add Team To Alert** Routes an alert to an additional team, which notifies them. The original responders keep it too. - **Add Responder To Alert** Adds a team, user, escalation, or schedule as a responder. The general form, and the way to pull an escalation policy onto an alert raised without one. - **Add Note To Alert** Appends a note to an alert's log. Where an automation records what it checked and what it found. - **Execute Alert Action** Runs one of the custom actions defined on the alert. The action has to have been declared when the alert was created or configured on its integration. - **Add Tags To Alert** Adds labels to an alert. Policies and saved searches filter on tags, so tagging is how a workflow reclassifies an alert into a different routing bucket. - **Remove Tags From Alert** Removes labels from an alert. - **Add Alert Details** Adds custom key and value properties to an alert. Where a workflow writes what it discovered so the next person reads it on the alert rather than in a log somewhere else. - **Remove Alert Details** Removes named custom properties from an alert. - **Update Alert Priority** Changes an alert's priority, which re-routes it, because notification and escalation policies branch on priority. - **Update Alert Message** Rewrites an alert's headline. It is what people see on their phone, so changing it mid incident changes what everyone is looking at. - **Update Alert Description** Replaces an alert's description. It replaces rather than appends, so adding context is usually a note instead. - **Delete Alert** Deletes an alert outright, taking its log with it. Closing is almost always what a workflow means. - **List Alert Logs** Returns an alert's activity log: every notification, action, and state change. Where an alert nobody answered explains itself. - **List Alert Notes** Returns the notes on an alert, which is the human commentary rather than the system log. - **List Alert Recipients** Returns who the alert was sent to and what state each recipient is in. The answer to whether anybody was actually notified. - **List Alert Attachments** Returns the files attached to an alert. - **Get Alert Attachment** Returns one attachment's metadata and a short lived download link. - **Delete Alert Attachment** Removes a file from an alert. ### Saved Searches 5 - **List Saved Searches** Returns the account's saved alert searches. A saved search keeps a complicated filter in one place so workflows do not repeat it. - **Get Saved Search** Returns one saved search, including its query. - **Create Saved Search** Saves an alert query under a name. It belongs to a real user rather than to the account. - **Update Saved Search** Changes a saved search's name, query, owner, or sharing. - **Delete Saved Search** Removes a saved search. Any workflow running it by name starts returning nothing rather than failing. ### Incidents 19 - **Create Incident** Declares an incident, the coordinated response above an alert. Asynchronous like alerts, and the service id is what puts it on a status page and starts impact tracking. - **Get Incident Request Status** Turns the request id from an asynchronous incident write into its outcome, including the resulting incident id. - **Get Incident** Returns one incident, addressed by id or by short number depending on the identifier type. - **List Incidents** Returns incidents filtered with Opsgenie's query language. - **Close Incident** Closes an incident, meaning the work is finished. A different state from resolved. - **Resolve Incident** Marks an incident resolved, meaning customer impact has stopped. This is the one that updates the status page, and the one to reach for before closing. - **Reopen Incident** Takes a closed or resolved incident back to open, which keeps the timeline and postmortem in one place rather than splitting a recurrence across two records. - **Add Note To Incident** Appends a note to an incident's log. - **Add Responder To Incident** Pulls a team, user, escalation, or schedule onto an incident, which notifies them. - **Add Tags To Incident** Adds labels to an incident. - **Remove Tags From Incident** Removes labels from an incident. - **Update Incident Priority** Changes an incident's severity. - **Update Incident Message** Rewrites an incident's headline. - **Update Incident Description** Replaces an incident's description. Adding context is a note instead. - **Get Incident Associated Alerts** Returns the ids of the alerts rolled up into an incident. How a workflow closes every alert an incident covered rather than leaving them open behind a resolved incident. - **Get Incident Responder Alerts** Returns the ids of the alerts raised to notify the incident's responders, which are separate records from the alerts that caused it. - **List Incident Logs** Returns an incident's activity log. - **List Incident Notes** Returns the notes on an incident. - **Delete Incident** Deletes an incident outright, along with its timeline and postmortem. ### Incident Timeline 7 - **List Incident Timeline Entries** Returns an incident's timeline, which is the curated narrative rather than the raw log, and what a postmortem is written from. - **Create Incident Timeline Entry** Adds an entry to an incident's timeline. The timestamp is when the thing happened rather than when it is being recorded, which matters when an agent reports something after the fact. - **Get Incident Timeline Entry** Returns one timeline entry. - **Update Incident Timeline Entry** Rewrites a timeline entry. - **Hide Incident Timeline Entry** Hides an entry without deleting it. The recoverable way to remove noise from a timeline a review will be written from. - **Unhide Incident Timeline Entry** Brings a hidden timeline entry back. - **Delete Incident Timeline Entry** Removes a timeline entry permanently. ### Postmortems 5 - **Create Postmortem** Starts a postmortem for an incident. Worth automating the moment an incident resolves, because a review created while the timeline is fresh is the difference between one that happens and one that does not. - **Get Postmortem** Returns one postmortem in full. - **Update Postmortem** Rewrites a postmortem's content and fields. - **Update Postmortem Status** Moves a postmortem between draft and published. Publishing makes it visible beyond its authors. - **Delete Postmortem** Removes a postmortem. ### On Call 4 - **Get On Calls** Returns who is currently on call for a schedule. The most useful read in the service, because it lets a workflow address the person actually responsible rather than a name in a template. - **Get All On Calls** Returns who is on call across every schedule in one call. - **Get Next On Calls** Returns who takes over a schedule next, and when. The handover question, and how a workflow warns the incoming person about something still open. - **Get Schedule Timeline** Expands a schedule's rotations into the actual shifts over a time span with overrides applied. The only operation that shows the real roster, and what a coverage check depends on. ### Schedules 5 - **List Schedules** Returns the account's on call schedules. - **Get Schedule** Returns one schedule. Its rotations are only included when expanded. - **Create Schedule** Creates an on call schedule. A schedule with no rotations covers nobody, and Opsgenie creates it happily. - **Update Schedule** Changes a schedule's name, timezone, owner, or enabled state. The reversible alternative to deleting one. - **Delete Schedule** Removes a schedule and its rotations. Anything routing to it stops notifying. ### Rotations 5 - **List Schedule Rotations** Returns a schedule's rotations, which are the rules deciding who is on when. - **Create Schedule Rotation** Adds a rotation to a schedule. The participant order is the rotation order, and the start date is read in the schedule's timezone rather than the caller's. - **Get Schedule Rotation** Returns one rotation. - **Update Schedule Rotation** Changes a rotation. Sending participants replaces the whole rota in the order given, which is what this page's human gate exists for. - **Delete Schedule Rotation** Removes a rotation from a schedule. The hours it covered have nobody on them unless another rotation covers them. ### Overrides 5 - **List Schedule Overrides** Returns a schedule's overrides, which are the temporary swaps sitting on top of its rotations. - **Create Schedule Override** Puts somebody else on call for a period without touching the rota. The operation for sickness, holiday, and a swapped shift, and worth giving an alias since that is the only way to change it later. - **Get Schedule Override** Returns one override, addressed by its alias. - **Update Schedule Override** Changes an override's dates or who is covering. It replaces rather than merges, so send the whole record. - **Delete Schedule Override** Removes an override, putting the underlying rotation back in charge. ### Teams 13 - **List Teams** Returns the account's teams. - **Get Team** Returns one team with its members. - **Create Team** Creates a team, optionally with members. A team is what alerts route to, so one created empty is a routing target that notifies nobody. - **Update Team** Changes a team's name, description, or membership. Sending members replaces the whole membership. - **Delete Team** Removes a team. Everything routing to it stops notifying, and the failure is silent. - **List Team Logs** Returns a team's audit log of membership and configuration changes. - **Add Team Member** Adds somebody to a team without touching the rest of the membership. - **Remove Team Member** Takes somebody out of a team. Their place in the team's on call rotation does not update itself. - **List Team Routing Rules** Returns a team's routing rules, which decide which schedule or escalation an alert reaching the team goes to. Usually where an alert that went to the wrong person is explained. - **Create Team Routing Rule** Adds a rule deciding where an alert reaching the team goes. The first matching rule wins, and a new rule lands last unless it is ordered. - **Update Team Routing Rule** Changes a routing rule. - **Change Team Routing Rule Order** Moves a routing rule up or down. The operation that actually changes routing when the rules are right and their order is not. - **Delete Team Routing Rule** Removes a routing rule. Alerts it used to catch fall through to the next matching rule. ### Team Roles 3 - **List Team Roles** Returns a team's custom roles and what each may do. - **Create Team Role** Adds a custom role within a team with its own rights. - **Delete Team Role** Removes a custom team role. Members holding it fall back to the team default, which usually grants less. ### Users 18 - **List Users** Returns the account's users. - **Get User** Returns one user, and with contacts expanded, whether they can actually be paged. - **Create User** Adds a user to the account. They cannot be paged until their contact methods are verified, so a new account routed alerts notifies nothing. - **Update User** Changes a user's name, role, timezone, or details. The username itself cannot be changed. - **Delete User** Removes a user. Check their rotations first, because deleting somebody who is in one leaves a hole nothing reports. - **List User Schedules** Returns the schedules a user appears in. The list to read before offboarding somebody. - **List User Teams** Returns the teams a user belongs to. - **List User Escalations** Returns the escalation policies a user appears in, which is the other place a departure leaves a hole. - **List User Forwarding Rules** Returns the rules forwarding this user's notifications elsewhere. One left in place after somebody returns from leave keeps sending their pages away. - **List User Contacts** Returns a user's notification methods and whether each is enabled. The answer to whether this person can actually be paged. - **Create User Contact** Adds a notification method to a user. A phone number has to be in the country and number form the API expects, and the wrong format is accepted and never rings. - **Enable User Contact** Turns a notification method back on. - **Disable User Contact** Silences one of a user's notification methods. Disabling somebody's only method makes them unpageable while leaving them in every rotation. - **Delete User Contact** Removes a notification method from a user. - **List User Notification Rules** Returns a user's own rules deciding how and when they are told about an alert. - **Enable User Notification Rule** Turns one of a user's notification rules back on. - **Disable User Notification Rule** Turns off one of a user's notification rules. The alert creation rule is the one that pages them. - **Copy Notification Rules To Users** Copies one user's notification rules onto others. The onboarding shortcut that gives a new joiner the team's agreed escalation behavior rather than the defaults. ### Escalations 5 - **List Escalations** Returns the account's escalation policies. - **Get Escalation** Returns one escalation policy with its rules. - **Create Escalation** Creates the chain of who is told next when nobody answers. Each rule carries its own delay and condition, and the condition is what makes it an escalation rather than a broadcast. - **Update Escalation** Changes an escalation policy. Sending rules replaces the whole chain. - **Delete Escalation** Removes an escalation policy. Routing rules pointing at it stop escalating, and Opsgenie does not report which ones were affected. ### Heartbeats 8 - **Ping Heartbeat** Tells Opsgenie a scheduled job is alive. The inverse of every other operation here, because Opsgenie raises an alert when this stops arriving. It belongs at the end of a scheduled workflow rather than the beginning. - **List Heartbeats** Returns the account's heartbeats and whether each is enabled and expired. - **Get Heartbeat** Returns one heartbeat, including whether it has expired, which means the ping stopped and the alert was raised. - **Create Heartbeat** Creates a dead man's switch. Set the interval longer than the job's worst run rather than its usual one, or a slow night pages somebody about a job that is working. - **Update Heartbeat** Changes a heartbeat's interval, alert settings, or enabled state. - **Enable Heartbeat** Starts a heartbeat monitoring again. Its interval restarts from now. - **Disable Heartbeat** Stops a heartbeat monitoring. The right move before planned downtime, and the thing to remember to undo, because a disabled heartbeat looks exactly like a healthy one in a list. - **Delete Heartbeat** Removes a heartbeat. Anything still pinging it starts failing visibly, which a disabled heartbeat does not. ### Services 5 - **List Services** Returns the account's services, which incidents attach to and status pages show. - **Get Service** Returns one service. - **Create Service** Creates a service. A service belongs to a team, so there is no account level service. - **Update Service** Changes a service's name or description. - **Delete Service** Removes a service. Incidents attached to it lose their impact record. ### Maintenance 6 - **List Maintenance** Returns the account's maintenance windows, which suppress alerting for a period. - **Get Maintenance** Returns one maintenance window and what it suppresses. - **Create Maintenance** Suppresses alerting for a period. What a deployment workflow should do before it starts, and it suppresses genuine failures as well as expected noise. - **Cancel Maintenance** Ends an active window immediately, restoring alerting. What a deployment workflow should do on its error path as well as on success. - **Change Maintenance End Date** Moves an active window's end. The right call when a deployment runs long, rather than stacking a second window on the first. - **Delete Maintenance** Removes a maintenance window and its record. For a scheduled window no longer needed, where cancelling is for an active one. ### Integrations 6 - **List Integrations** Returns the account's integrations, which are the inbound channels alerts arrive through. - **Get Integration** Returns one integration. - **Create Integration** Creates an inbound integration and returns its API key. The key appears in that response and nowhere afterwards, so a workflow that discards it has created a channel it cannot use. - **Enable Integration** Lets an integration accept alerts again. - **Disable Integration** Stops an integration accepting alerts. Alerts sent to it are dropped rather than queued, so the sender reports success and pages nobody. - **Delete Integration** Removes an integration and invalidates its key. Whatever was posting to it starts failing visibly. ### Policies 7 - **List Alert Policies** Returns the alert policies for a team or the account, which rewrite or suppress alerts as they arrive. Often where an alert that looks wrong came from. - **List Notification Policies** Returns the notification policies deciding whether and how an alert actually reaches somebody. A de-duplication or delay policy here is a common reason a real alert arrives late. - **Get Policy** Returns one alert or notification policy with its filter and its actions. - **Enable Policy** Turns a policy back on. - **Disable Policy** Turns a policy off. The fastest reversible fix when a policy is suppressing something it should not during an incident. - **Change Policy Order** Moves a policy up or down the evaluation order, which decides whether a later policy ever sees the alert. - **Delete Policy** Removes a policy. Disabling is the reversible version and usually the right one mid incident. ### Account 1 - **Get Account** Returns the account's name, plan, and user count. The cheapest connection check, and the fastest way to confirm the key and the region match. ## Frequently Asked Questions ### What can FlowRunner do with Atlassian Opsgenie? FlowRunner agents can run Create Alert, Get Alert Request Status, and Get Alert in Atlassian Opsgenie, plus 153 more actions. ### Does connecting Atlassian Opsgenie to FlowRunner require OAuth? No. Atlassian Opsgenie connects to FlowRunner with an API key, no OAuth flow required. ### Can Atlassian Opsgenie trigger a FlowRunner workflow automatically? Atlassian Opsgenie doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger. **Work at Atlassian Opsgenie?** This integration exposes Atlassian Opsgenie to AI agents on every FlowRunner plan, including through MCP, at no cost to you. [See what FlowRunner offers integration partners](https://flowrunner.ai/integrations/partners), including how to keep this page current. --- Markdown version of https://flowrunner.ai/integrations/opsgenie. Site index: https://flowrunner.ai/llms.txt