Did something suddenly get more expensive?
Each service's recent daily cost is compared against its own recent history, so a jump that's unusual for that service gets flagged — even if the absolute amount is small.
1 min read
How it works
What the Cost Anomalies tab shows.
- Baseline vs. recent
- A service is flagged when its recent average daily cost clears a small floor and comes in well above its own usual range — not against one fixed threshold shared by every service.
- Insufficient data
- Shown for services that haven't been running long enough in the selected window to have a reliable baseline yet — it means "too early to tell", not "nothing changed".
- Monthly impact
- The extra daily cost of the anomaly, projected over 30 days, so you can size how much the jump matters before digging in.
- Alerts and sensitivity
- Owners and admins are alerted through the channels they enabled, once per episode. Sensitivity and the minimum impact to alert are adjustable in the organization settings — see Notifications.
Reading the results
| Label | What it means |
|---|---|
| Flagged as anomalous | This service's recent daily cost is well above what's normal for it, based on its own history — not a shared threshold applied to every service alike. |
| Insufficient data | This service hasn't been running long enough in the selected window to establish a reliable baseline — it's "too early to tell", not "nothing changed". |
| Monthly impact | The extra daily cost of the anomaly, projected over 30 days — a quick way to size whether it's worth investigating now. |