Mcp Observability Server
Query production logs and traces from observability backends through MCP so your agent debugs incidents without opening five vendor UIs.
Overview
MCP Observability Server is an Operate-phase MCP server that lets agents query logs and traces from observability platforms including New Relic via unified MCP tools.
What is this MCP server?
- PyPI mcp-observability-server v0.1.3 with stdio transport
- Unified log query and tracing across observability platforms from one MCP surface
- New Relic integration via NEW_RELIC_API_KEY and NEW_RELIC_ACCOUNT_ID env vars
- Agent-driven search instead of manual NRQL or filter UI hopping
- GitHub gagandeeppra/mcp-observability-server for extension and self-host
- PyPI mcp-observability-server version 0.1.3
- Stdio transport documented in server manifest
- 2 New Relic environment variables listed in server.json
What problem does it solve?
During outages you copy paste log snippets into chat because your agent cannot reach the same search and trace views you use in New Relic or similar tools.
Who is it for?
Solo operators on New Relic (or compatible backends) who want agent-assisted log and trace investigation during on-call style debugging.
Skip if: Greenfield projects with no observability vendor, or teams needing full APM agent installation instead of API-backed query tools.
What do I get? / Deliverables
Registered MCP tools let the agent run observability queries with your API credentials so diagnosis stays inside the coding session.
- MCP tools for cross-platform log query and trace exploration
- New Relic-backed investigation from the agent workspace
- Stdio MCP config for operate-time monitoring assistance
Recommended MCP Servers
Journey fit
After ship, solo operators live in logs and traces—Operate is where unified observability access pays off daily. Monitoring is the canonical shelf for log search and tracing across platforms, not feature development or marketing.
How it compares
Observability query MCP—not a log shipper, dashboard builder, or PagerDuty replacement.
Common Questions / FAQ
Who is MCP Observability Server for?
Builders who already send logs and traces to supported platforms and want MCP access for AI-assisted incident search.
When should I use MCP Observability Server?
In Operate monitoring when you need fast log and trace lookups while fixing errors or validating a deploy without leaving your agent IDE.
How do I add MCP Observability Server to my agent?
Install mcp-observability-server from PyPI, set NEW_RELIC_API_KEY and NEW_RELIC_ACCOUNT_ID (and any other backend vars your install needs), then register stdio in MCP settings.