Alerts notify you when important metrics cross thresholds, so you can respond quickly to issues.Documentation Index
Fetch the complete documentation index at: https://docs.duckie.ai/llms.txt
Use this file to discover all available pages before exploring further.
What are Alerts?
Alerts are automated notifications triggered by metric conditions:- Escalation rate exceeds threshold
- Response time degrades
- Error rate spikes
- Volume anomaly detected
Creating Alerts
Select Metric
Choose what to monitor:
- Escalation rate
- Resolution rate
- Response time
- Error rate
- Volume
Set Condition
Define when to trigger:
- Threshold value
- Duration (sustain for X minutes)
- Comparison (above, below, change)
Alert Conditions
Threshold Alerts
Trigger when a metric crosses a value:| Condition | Example |
|---|---|
| Above | Escalation rate > 25% |
| Below | Resolution rate < 70% |
| Change | Volume increased > 50% |
Duration
Require the condition to persist:Scope
Apply to specific deployments:- All deployments
- Specific deployment
- Specific channel
Example Alerts
High Escalation Rate
| Setting | Value |
|---|---|
| Name | High Escalation Alert |
| Metric | Escalation rate |
| Condition | Above 25% |
| Duration | 10 minutes |
| Notification | Slack #support-alerts |
Degraded Response Time
| Setting | Value |
|---|---|
| Name | Slow Response Alert |
| Metric | Response time P95 |
| Condition | Above 30 seconds |
| Duration | 5 minutes |
| Notification | Email + Slack |
Volume Spike
| Setting | Value |
|---|---|
| Name | Volume Anomaly |
| Metric | Conversation volume |
| Condition | 2x normal for time of day |
| Notification | Slack #support-team |
Resolution Drop
| Setting | Value |
|---|---|
| Name | Low Resolution Alert |
| Metric | Resolution rate |
| Condition | Below 60% |
| Duration | 1 hour |
| Notification |
Notification Channels
- Send to specific addresses
- Include alert details and links
- Configure frequency (immediate, digest)
Slack
- Post to a channel
- Direct message
- Include actionable links
Webhook
- POST to your endpoint
- Include alert data as JSON
- For custom integrations
Managing Alerts
Active Alerts
View currently triggered alerts:- What triggered
- When it started
- Current value
Alert History
See past alerts:- When triggered
- How long it lasted
- When resolved
Edit/Disable
Modify existing alerts:- Click on the alert
- Edit settings
- Save changes
- Toggle status to disabled
- Alert won’t trigger
- Re-enable when ready
Responding to Alerts
When an alert triggers:Best Practices
Start Simple
Begin with critical alerts only:- High escalation rate
- Error spikes
- Major volume anomalies
Set Appropriate Thresholds
Avoid alert fatigue:- Too sensitive = too many alerts
- Too loose = miss real issues
- Start conservative, adjust based on experience
Use Duration
Require sustained conditions:- Prevents alerting on momentary blips
- Focus on real problems
- Reduce noise
Route to Right People
Send alerts where they’ll be acted on:- Operations team for volume spikes
- Engineering for error rates
- Support leads for escalation issues
Next Steps
Performance Metrics
Understand metrics to alert on
Run History
Investigate alert causes