Insights
Insights surfaces automatically detected patterns, anomalies, and recommendations from your platform data — so you don't have to manually spot trends in dashboards.
What Insights detects
| Insight type | Example |
|---|---|
| Traffic anomalies | "Request volume to POST /v1/events is 3× higher than your 7-day average" |
| Error spikes | "Error rate for workflow Invoice Approval increased from 0.4% to 8.2% in the last hour" |
| Unused resources | "API key ci-pipeline has not been used in 60 days" |
| Rate limit proximity | "At current rate, you will hit your monthly request limit in ~6 days" |
| Latency regression | "p99 latency on GET /v1/workflows increased 40ms vs. last week" |
| Workflow efficiency | "Step send_notification accounts for 78% of total workflow run time" |
How insights are delivered
- Dashboard panel — the Insights panel on your home page shows the top 5 active insights, refreshed hourly
- Notifications — configure critical insights to trigger an alert via email, Slack, or PagerDuty
- API — query current insights programmatically:
GET /v1/insights
Dismissing and snoozing
Insights can be dismissed (won't reappear for the same pattern) or snoozed for a set period. If the underlying condition changes significantly, a dismissed insight will resurface automatically.
Insight-driven automations
Use insights as triggers for automations. For example, when a "rate limit proximity" insight fires, automatically notify your team and trigger a workflow to review and prune unused integrations — closing the loop from detection to resolution automatically.