Section 8 · 10 min
Starting an Incident Yourself
Not every incident starts with an alert
Sometimes you know about a problem before monitoring does. You should be able to start the operational process without waiting for a signal.
When you raise one manually
- A customer reports an outage before any alert fires
- You observe degradation directly
- Another team escalates verbally
- A change goes wrong
- Monitoring did not catch it
The manual path, before and after
How it used to work
- Agent
- AWS Console
- Find Script
- Execute
- Integration
- Opsgenie
- Incident
How it works in JSM
- Agent
- JSM Portal
- Service Offering
- Incident — Standard or Major
- Responders / On-Call
- Coordination
- Resolution / PIR
You no longer need to know the technical mechanism that creates an incident — you raise it from the portal and the response process takes over.
You initiate the operational process. You do not need to know the technical mechanism behind it — choose the right service offering, state clearly what is happening, and pick standard or major based on how much coordination and communication is required.
Key Concept 8
If you know service is impacted, start the incident — do not wait for an alert to confirm it.
What's new for responders
Alert grouping
Related alerts may be grouped together to reduce noise and improve situational awareness. Twenty alerts describing one outage should not become twenty separate operational conversations.
- 20 Related Alerts
- Alert Group
- Understand One Operational Event
More connected Jira work
JSM brings alerts, incidents, services, responders and Jira work items closer together, so operational work and engineering work stop living in separate worlds.
- Alert
- Incident
- Service
- Jira work item
More service context
Affected services help responders understand what business or technical capability is actually impacted. That becomes particularly useful when many technical alerts contribute to one incident.
- Alert
- Affected Service
- Business Impact
Twenty alerts, one event
When many alerts describe the same outage, treat them as one operational event. Grouping and related alerts exist so the response stays a single conversation.
Live demo
- Open the JSM portal
- Choose the service offering
- Describe what is happening
- Select standard or major
- Submit
- Open the resulting incident