Retired service

Amazon Lookout for Metrics is gone. Keep anomaly detection simple.

AWS ended support for Lookout for Metrics on October 10, 2025. If you only need to send a time series and get anomalies, expected values, and bands back, Go Insights is the smaller replacement path.

detector.response
{
  "period": 7,
  "expectedValues": [812, 798, 805, 834],
  "upperMargins": [38, 36, 37, 41],
  "isAnomaly": [false, false, false, true],
  "competitor": "AWS Lookout",
  "replacement": "Go Insights"
}
Status AWS ended support on Oct 10, 2025
Best fit Teams replacing retired AWS anomaly jobs
Go Insights angle Focused API, predictable usage pricing
Short answer

Choose based on the job, not the logo.

Lookout for Metrics was a managed AWS anomaly service. Go Insights is narrower: a JSON API for scoring time series without rebuilding the detector inside your own stack.

Comparison

Go Insights vs Amazon Lookout for Metrics

Question Amazon Lookout for Metrics Go Insights
Service status Support ended; no new signups before retirement. Available as a hosted API and live demo.
Implementation shape AWS-native managed service with migration now required. Send JSON time series, receive expected values, bands, period, and anomaly flags.
Best for Teams already committed to rebuilding on AWS services. Teams that want a smaller replacement instead of a full ML rebuild.
Pricing posture Retired, so price is no longer the main issue. Free test tier, paid plans from $25/month, overage from $0.20/1K.

Use AWS Lookout if...

  • You want to rebuild the old workflow inside AWS using SageMaker, CloudWatch, Lambda, or custom services.
  • Your data platform and deployment process are already locked into AWS-only operations.

Use Go Insights if...

  • You want a hosted detector that behaves like an API, not a new ML project.
  • You need expected ranges and anomaly flags that can be shown in your own product or alerting flow.
  • You want a free tier, a small paid plan, and simple overage pricing.
Migration path

Test with the payload shape first.

Use the live tester before committing engineering time. The demo uses the same kind of request and response fields your app would consume in production.

  1. Start with one retired Lookout job and send the same historical time series to the Go Insights API demo.
  2. Compare anomaly points, expected values, and period detection before wiring alerts.
  3. Move production traffic behind your own API key once the response shape matches the workflow.
Pricing

Lower commitment before usage proves itself.

Paid plans start at $25/month. Overage starts at $0.20 per 1,000 detections, then falls on larger plans. The larger plans include more detections and reduce the overage rate to $0.12 per 1,000 detections.

Bring your own time-series.

Run one real metric through the tester, inspect the JSON, and decide whether the API fits the replacement job.

Sources

Comparison claims use public vendor documentation and public pricing pages. Links open in a new tab.