Configuration Elements
An escalation rule contains four core elements. The system matches rules from top to bottom, stopping after the first successful match.1. Trigger Conditions
Determines which incidents trigger the current rule. Effective Time: Controls when the rule participates in matching.- Off (Default): Active 24/7
- On: Only participates in matching within specified time ranges; incidents outside this time will skip this rule. Two modes supported:
- Specific Time Period: e.g., only active on weekdays (Monday-Friday 9:00-18:00)
- Service Calendar: Reference a predefined Service Calendar, e.g., only active on “Trading Days”, automatically excluding holidays
- Off (Default): All incidents may match this rule
- On: Precisely filter by incident attributes, supports matching by title, severity (e.g., Critical only), labels (e.g.,
service=payment), etc.
2. Notification Recipients
Determines who receives notifications.| Recipient Type | Description |
|---|---|
| Schedule (Recommended) | Assign to currently on-call personnel |
| Team | Assign to all members of a team |
| Individual | Assign to specific personnel |
| Combination Mode | You can select multiple types simultaneously |
3. Notification Methods
Determines how users are reached.- Direct Message (Reach Individuals)
- Group Chat (Reach Groups)
Point-to-point messaging to specific assignees, supports phone calls, SMS, email, app push, IM direct messages (Feishu/Lark, Dingtalk, WeCom).
- Follow Personal Preferences (Recommended): Notification methods configured by members in Personal Settings
- Follow Unified Settings: Enforce notification methods (e.g., must send SMS), overriding member personal preferences
4. Escalation Rules
This is the key mechanism for ensuring incident closure. When first-level responders don’t respond or complete handling, the system automatically escalates to the next level. Escalation Conditions:- Timeout without acknowledgment: No one acknowledges within N minutes after incident trigger
- Timeout without closure: Incident not resolved within N minutes after trigger
| Scenario | Escalation Path |
|---|---|
| Level 1 → Level 2 | Level 1 on-call (SRE) no response for 15 min → Escalate to Level 2 development experts |
| Primary/Backup Escalation | Primary on-call no response for 10 min → Escalate to backup on-call |
| Tiered Reporting | Technical staff unresolved for 30 min → Escalate to Tech Lead → Escalate to CTO |
Best Practices
Fallback Policy
Fallback Policy
Recommend keeping a default policy without filter conditions at the bottom of the list, assigned to the SRE team or administrators, to prevent incidents from being missed due to filter rule misconfigurations.
Avoid Over-notification
Avoid Over-notification
- For low-priority
Info/Warninglevel alerts, recommend using only IM messages, not phone calls or SMS - Use the Loop Notification feature cautiously to avoid alert bombardment
Reasonable Channel Splitting
Reasonable Channel Splitting
Don’t try to manage all company alerts in one channel. Splitting channels by functional module or team, with each channel maintaining independent and simple escalation rules, is key to reducing maintenance costs.
FAQ
Alerts are generated but no notifications received?
Alerts are generated but no notifications received?
Go to Incident Details → Timeline to check if notification status is normal for each channel. If there are failure messages, use them for troubleshooting. Contact technical support for more issues.
Notification method doesn't match personal preference settings?
Notification method doesn't match personal preference settings?
Flashduty On-call direct messaging supports two modes: “Follow Personal Preferences” and “Follow Unified Settings”. Only in “Follow Personal Preferences” mode will the system send notifications according to your personal settings.Go to Channel Details → Escalation Rules to check current settings.