About the service
Observability built on OpenTelemetry from day one: no proprietary agent, no format lock-in — it works with any OTel-compatible tool, agent or SDK immediately. The foundation is SignalStore, which holds logs, metrics, traces, profiles, events and real-user data; you query it with PromQL and SQL, through a CLI, through an API and through an MCP server, so both humans and AI agents can work the telemetry. Three products sit on top: observability (application performance monitoring, real-user monitoring, infrastructure, and language models and agents as a separate view), Agent0 for automation and fixing (live insights, root cause analysis, automatic fixes and auto-tuning), and Darkplane, the control room for what AI writes in your codebase. There is Kubernetes monitoring, distributed tracing, log management, website and synthetic monitoring, dashboards, alerting and service maps; dashboards and alerts are defined as code, and the interface is built for keyboard-first navigation. A particular strength is cost control: you pay only for the telemetry you send, with no per-seat and no per-gigabyte-stored fees, and the platform has monthly budget limits, filters for noisy data, spend forecasting and a hard cap on the invoice with a warning and a grace period before ingestion pauses. Every capability is included in a single plan — the vendor has no cut-down feature tiers.