RunLLM

RunLLM is a web-based AI SRE (Site Reliability Engineering) tool for SREs, support engineers, and technical teams that need faster incident investigation across complex observability stacks.

It connects to telemetry and operational context to investigate alerts automatically and produce actionable outputs:

• Rapid RCA and mitigation guidance by correlating alerts with logs, metrics, traces, tickets, code, and documentation, then returning prioritized next steps with verification checks.
• Self-teaching investigations that map a team’s stack, learn from incidents and user corrections, and reuse proven investigation steps over time.
• Agentic workflows and customization (v2) with a UI for managing suites of agents plus a Python SDK for custom triage, escalation, and support workflows.
• Connectors and integrations including Datadog, Splunk, GCP Logging, Grafana, Slack, and Zendesk, with read-only by default, full action logging, and SOC 2 Type II compliance.

Pricing: Contact for pricing.

Notable for combining incident-response automation with multi-agent collaboration, validated code execution, and real-time reasoning, and for being used by vLLM to deflect 99% of 13k+ monthly technical questions.