Monitoring & Observability

Multi-agent observability: cascade failure detection, heartbeats, and forensic replay

MCP server for 402 Index — search 15,000+ paid API endpoints across L402, x402, and MPP protocols with real-time health monitoring

PumpFun on-chain intelligence — Telegram feed + MCP server for AI assistants. Query token data, GitHub social fee claims, creator profiles, and more.

MCP server that diagnoses pg-boss job queues — evidence-backed findings for retry storms, stuck jobs, missed schedules and expiry overruns, each with the safest recovery action.

One firewall and password manager for all your AI coding agents' MCP servers and secrets.

MCP server that gives agents ground truth of what is ACTUALLY running on this machine — agent sessions, listening ports, daemons — read live from the OS, never from stale docs or memory.

MCP server that collects telemetry from any mcp-telemetry-sdk-instrumented tool and pushes live progress to any connected MCP client (Claude Code, Cursor, and others).

Orchestration + observability + coach layer for Claude Code token optimization. Measures tool usage, enforces budgets, advises on complementary tools (serena, RTK), and proactively surfaces savings tips.

AI agent vision for GitHub repository monitoring — with ROI measurement. perceptdot.com

MCP-сервер поверх INSO Error Logs Service. Read-only tools (search, analytics, fingerprinting, correlations) для подключения Claude Desktop/Code и ChatGPT к логам ошибок. Streamable HTTP transport + stdio для npx-запуска.

MCP server for Linux system monitoring — CPU, memory, disk, network, processes, systemd services, and system logs for Claude, Cursor, and AI agents

CodeAtlas MCP server — 57 tools + 8 resources that expose live codebase architecture (routes, sequences, diffs, impact analysis, architecture violations, read-only SQL access) and AI-driven code review (per-entry-point review, evidence-gated findings, cus

A Model Context Protocol (MCP) server that provides intelligent web reading capabilities using the Jina AI Reader API. It extracts clean, LLM-ready content from any URL.

MCP server that scaffolds Chassis backends — an Express 5 + TypeScript starter with a database, auth and an optional Next.js front end. Lets an AI agent create a working, tested project in one tool call.

perceptdot MCP server for Sentry error tracking

MCP server providing AI agent observability — tracing, cost tracking, performance monitoring, and audit trails

Outage radar for AI agents. LLM pricing, status, inference performance, and real-time agent-reported failure signals. 9 tools. No auth required.

Step-by-step observability for MCP agent workflows

Health monitoring for all your MCP servers

Ionhour MCP Server -- connect AI assistants to your uptime monitoring workspace