Alerting overview

FireTail's alerting system allows you to monitor and respond to specific conditions or anomalies within your AI environment. Alerts are designed to notify you when thresholds you set up are met, helping to manage and mitigate potential issues effectively.

Alert types

• Static alerts: Allow monitoring of AI-specific behaviors (e.g., token usage, model latency) using fixed thresholds.
• Anomaly alerts: Use historical data to detect unusual activity. These alerts detect irregular patterns that fall outside of established data ranges. Anomaly alerts can be adjusted for sensitivity to detect smaller or larger deviations.

Notification integrations

Alerts are linked to notification integrations. These function as the communication method for the alert. Notifications can be sent through various channels, including HTTP webhooks, Slack, Teams, PagerDuty, Jira, and more.