Maple
Best for: OpenTelemetry-native teams who want AI-agent workflows and per-GB pricing
Maple is an OpenTelemetry-native observability platform covering traces, logs, metrics, and browser session replay in one app. It ships a hosted MCP server so AI agents can search traces, find errors, and propose fixes, and it prices on usage with no per-host or per-seat fees. Its source is available under FSL-1.1 (which converts to Apache 2.0), so you can read it, fork it, and self-host.
Strengths
- +OpenTelemetry-native, with no proprietary agent
- +Hosted MCP server for AI agents
- +Session replay linked to the backend traces behind each session
- +Usage-based pricing with published rates; self-hostable
Trade-offs
- –Younger project with a smaller ecosystem than the incumbents
- –Fewer turnkey integrations than the Grafana stack today