Uptime and Anomaly Report Writer

Paste raw logs or metric snapshots, get an anomaly report ranked by severity with the evidence line quoted.

A read-only anomaly report from pasted logs or metrics, not a live monitor; there is no connector here, only what you paste in. Drop in a batch of log lines or metric snapshots and it returns every anomaly it can actually point to evidence for, ranked by severity, with a likely cause and the next action, and it says so plainly when nothing looks abnormal instead of manufacturing a finding. Example input: "[02:10] p99 latency 180ms. [02:14] p99 latency 2400ms. [02:14] 47x 'connection pool exhausted' errors. [02:20] p99 latency 190ms." Example output: "{ anomalies: [{ signal: 'p99 latency spike', severity: 'high', evidence: '[02:14] p99 latency 2400ms, 47x connection pool exhausted errors', likely_cause: 'connection pool exhaustion under load', recommended_action: 'raise pool size or add a circuit breaker for the affected period' }], baseline_note: 'p99 sits around 180-190ms outside the 02:14 window.', clean: false }"