Robel launched LogClaw, an open-source AI SRE platform that automatically creates Jira tickets from Kubernetes logs. The system ingests logs via OpenTelemetry and uses signal-based composite scoring to detect anomalies like OOMs, crashes, and dependency failures. Critical issues trigger detection in under 100ms, achieving 99.8% accuracy for high-severity failures while filtering out noise like validation errors and 404s.

Once an anomaly is confirmed, LogClaw's 5-layer trace correlation engine reconstructs the blast radius across services and generates root cause analysis. The platform then creates deduplicated tickets in Jira, ServiceNow, PagerDuty, OpsGenie, Slack, or Zammad within about 90 seconds. The architecture uses OTel Collector → Kafka → Python bridge → OpenSearch, with support for OpenAI, Claude, or Ollama for fully air-gapped deployments.

The managed cloud version costs $0.30/GB ingested, offering significant savings compared to Splunk ($1.2M/year at 500GB/day) and Datadog ($509K/year). Licensed under Apache 2.0, LogClaw addresses enterprise compliance needs by keeping logs within the VPC and eliminating third-party data exposure risks.