Skip to main content

​
Usage

langsight mcp-health [OPTIONS]

​
Options

OptionDefaultDescription
--configauto-discoveredPath to .langsight.yaml
--jsonfalseOutput results as JSON

​
What it does

Connects to each configured MCP server, runs initialize + tools/list, measures latency, and detects schema drift vs. the last stored snapshot. Exit codes:
  • 0 — all servers healthy
  • 1 — one or more servers DOWN or DEGRADED

​
Example output

MCP Server Health  (6 servers)
──────────────────────────────────────────────────────────────────
Server           Status      Latency   Tools   Schema
snowflake-mcp    ✓ up        142ms     8       bcf0ec26…
github-mcp       ✓ up        89ms      12      d2125e3a…
slack-mcp        ⚠ degraded  1,240ms   4       —         schema drift detected
jira-mcp         ✗ down      —         —       —         timeout after 5s
postgres-mcp     ✓ up        31ms      5       f4a2b1c9…
filesystem-mcp   ✓ up        12ms      6       8e3d7f2a…

4/6 servers healthy

​
Schema drift detection

LangSight stores a hash of each server’s tool list on the first run. On subsequent runs, if the hash changes, the server is marked DEGRADED and a warning appears. This catches:
  • Unplanned tool additions or removals
  • Changed tool descriptions (potential poisoning)
  • Breaking parameter changes

​
JSON output

langsight mcp-health --json | jq '.[].status'

[
  {
    "server_name": "postgres-mcp",
    "status": "up",
    "latency_ms": 31.4,
    "tools_count": 5,
    "schema_hash": "f4a2b1c9...",
    "checked_at": "2026-03-17T12:00:00Z"
  }
]

​
Data persistence

Results are stored in ~/.langsight/data.db (SQLite, default) or ClickHouse (production). Use --config to point at a different .langsight.yaml with a different storage backend.