AI
Anomaly detection, forecasts, natural-language query, and root-cause analysis.
The AI page in the console calls the Sensobit AI engine on the same namespace you are viewing. Daily limits depend on the billing plan.
Capabilities
| Feature | What it does |
|---|---|
| Anomaly detection | Flags unusual metric shapes with a confidence score |
| Anomaly alerts | Optionally pages you when a model fires |
| Forecasting | Projects CPU, memory, disk, and custom series forward |
| Natural-language query | Turns a sentence into a metric or log lookup |
| Root-cause analysis | Correlates metrics, logs, and traces around an incident |
| Remediation | Suggests (or, on higher plans, applies) a runbook step |
| Alert priority | Ranks open alerts so on-call sees the most likely cause first |
Example questions
- "Which hosts ran above 80% CPU in the last six hours?"
- "Show error logs for checkout after 04:00 UTC."
- "Why did latency jump on web-01?"
The engine only sees data in the active namespace. It cannot query another team's hosts.
API
The same features are exposed under /api/v1/anomaly, /api/v1/predictive, /api/v1/nlp, /api/v1/root-cause, /api/v1/remediation, /api/v1/alert-priority, and /api/v1/feedback on api.sensobit.com. Try a request in the playground.
Feedback you submit from the console is stored and used to tune ranking, not to train on raw customer logs outside your workspace.