alertloop reads your real server logs and names the likely cause, so you start with an answer instead of a "something is down" alert.
AI root cause analysis uses a language model to read the logs and signals around an incident and explain the likely reason it happened. In alertloop, every incident gets an automatic write-up covering services, Docker containers, pm2 and systemd processes, websites, and sustained CPU or memory spikes.
When the analysis is ready, the full root cause is emailed to you, with one-click approval for any suggested fix.
Ask about an incident in plain language and get answers grounded in your own telemetry.
OpenAI, Anthropic Claude, Google Gemini, Groq or self-hosted Ollama, with separate models for analysis and chat.
Per-account token limits, alerts at 80% and 100%, an optional hard cap, and a per-call cost log.
A service, container or site goes down, or CPU or memory stays high.
alertloop analyzes your real logs for that service.
The write-up arrives with any suggested fix for your approval.
Related: Docker monitoring · pm2 monitoring · systemd monitoring