MCP server category
Monitoring and observability MCP servers: comparison and full list
The Monitoring and observability category lists 57 open-source MCP servers, checked on GitHub. The best scored on our grid are Netdata, MCP grafana, and Openstatus. 77% received a commit in the last three months.
What Monitoring and observability MCP servers are for
Monitoring MCP servers give an AI agent access to information used to observe systems and investigate incidents. They serve operations teams, developers and people responsible for service availability. Depending on the integration, you can query metrics, inspect logs or manage incident tracking resources. Netdata describes exploration and root cause analysis across its monitoring data. MCP grafana supports dashboard search and queries against data sources in your Grafana instance. A practical request might be: “Review the available evidence around this incident and prepare a list of explanations for the team to investigate.” Results depend on the data the agent can access.
Example uses:
- Query available Grafana data sources to investigate a service degradation reported by your team.
- Explore Netdata metrics and logs associated with a service showing unusual operational behavior.
- Prepare a summary of incidents and maintenance events recorded in your Openstatus workspace.
- Find the owner of a data asset by searching the Marmot Data Catalog.
The category in figures
Of the 57 servers in the category (2% of the directory), 12 are official vendor implementations, 95% declare a licence, 11 offer a remote server and 14 need at least one API key or token according to the MCP registry. The median repository has 24 stars.
All Monitoring and observability MCP servers
Click a header to sort. Score out of 100: adoption, maintenance and framework (licence, registry, official status); risk: system access, secrets, licence, abandonment.
| Server | Stars | Last commit | Licence | Risk | Score |
|---|---|---|---|---|---|
| Netdata ◆ Discovery, exploration, reporting and root cause analysis over Netdata metrics, logs, systems, containers, processes and network connections | 81,115 | Oct 24, 2026 | GPL-3.0 | low | 100 |
| Openstatus Manage monitors, status pages, incidents and maintenances in your openstatus workspace. | 9,221 | Oct 24, 2026 | AGPL-3.0 | low | 91 |
| OneUptime Manage incidents, alerts, monitors, status pages, on-call and telemetry in OneUptime. | 7,754 | Oct 24, 2026 | — | moderate | 80 |
| MCP grafana ◆ Search dashboards, investigate incidents and query datasources in your Grafana instance. | 3,598 | Oct 23, 2026 | Apache-2.0 | moderate | 96 |
| Marmot Data Catalog Open-source data catalog. Search assets, explore lineage, and find ownership. | 623 | Oct 24, 2026 | MIT | low | 73 |
| Watch skill Watch video and live sessions, keep timestamped evidence, and verify an agent's own work. | 520 | Oct 14, 2026 | MIT | low | 72 |
| Homebutler Homelab management across multiple servers via SSH: system resource monitoring, Docker containers, Wake-on-LAN, network scans, open-port che | 305 | Oct 17, 2026 | MIT | low | 75 |
| Zabbix MCP server Zabbix integration for hosts, items, triggers, templates, problems, data and more. | 261 | May 10, 2026 | GPL-3.0 | low | 57 |
| Lighthouse MCP MCP server for Google Lighthouse performance metrics | 212 | Oct 21, 2026 | MIT | low | 68 |
| Dynatrace Connect AI assistants to Dynatrace to query observability data and analyze problems. | 167 | Oct 23, 2026 | Apache-2.0 | low | 67 |
| Logfire MCP ◆ Access OpenTelemetry traces and metrics through Logfire. | 161 | Jul 13, 2026 | MIT | low | 65 |
| GoldenAnalysis Read-only cross-cutting analysis, metrics, and reporting across the Golden Suite. | 133 | Oct 24, 2026 | MIT | low | 66 |
| SigNoz MCP Server SigNoz MCP server: access metrics, traces, logs, alerts, and dashboards via AI assistants | 128 | Oct 24, 2026 | Apache-2.0 | low | 66 |
| TraceWeave Portable MCP runtime for RTL logs and VCD; repository setup adds FSDB and site EDA integration | 121 | Oct 13, 2026 | MIT | low | 66 |
| Langfuse MCP Query Langfuse traces, debug exceptions, analyze sessions and manage prompts for LLM applications. | 112 | Oct 24, 2026 | MIT | moderate | 70 |
| MCP monitor Real-time system metrics: CPU, memory, disk, network, host and process information. | 91 | Aug 2, 2025 | Apache-2.0 | moderate | 37 |
| Lucidity MCP Improve AI-generated code quality with prompt-based analysis across 10 dimensions, from complexity to security vulnerabilities. | 90 | Mar 19, 2025 | Apache-2.0 | moderate | 37 |
| Pagespeed insights MCP Google PageSpeed Insights, CrUX and Lighthouse analysis: page comparison, real-user data, batch runs, regression tracking, diagnostics and r | 76 | Oct 21, 2026 | Apache-2.0 | moderate | 69 |
| MCPSpend Real-time cost observability for MCP tool calls: a transparent proxy that attributes spend per tool, project and end customer. | 66 | Aug 31, 2026 | MIT | moderate | 63 |
| Thoughtbox Extended reasoning with structured processes, persistence, and workflow guidance | 65 | Jul 15, 2026 | MIT | low | 51 |
| Last9 MCP server Bring real-time production logs, metrics and traces from Last9 into your local environment to fix code. | 63 | Oct 24, 2026 | Apache-2.0 | low | 63 |
| PIO Agent Build, inspect, flash, and monitor PlatformIO firmware projects with policy-controlled tools. | 54 | Oct 13, 2026 | MIT | low | 62 |
| Metoro MCP server ◆ Query and interact with Kubernetes environments monitored by Metoro. | 51 | Jun 2, 2026 | MIT | low | 60 |
| Raygun Crash reporting and real user monitoring via the Raygun API. | 30 | Sep 2, 2026 | — | moderate | 50 |
| Grafana loki MCP Query Loki logs through the Grafana API. | 29 | Jan 5, 2026 | MIT | low | 40 |
| Llmxray Ask what went wrong with your own Ollama calls: lost prompt cache and the token where the prompt stopped matching, context overflow, a model | 28 | Oct 24, 2026 | Apache-2.0 | low | 64 |
| Ig MCP server Debug container and Kubernetes workloads with eBPF via Inspektor Gadget. | 28 | Aug 13, 2026 | Apache-2.0 | low | 54 |
| SolidPing ◆ Uptime monitoring over MCP. List and create checks, read incidents and results, manage maintenance windows and status pages. Self-hosted sin | 26 | Oct 24, 2026 | AGPL-3.0 | low | 74 |
| AgentLens Tamper-evident AI agent observability: hash-chained audit log with verification and signed export (EU AI Act Art. 12), plus OpenTelemetry Ge | 24 | Oct 18, 2026 | MIT | low | 64 |
| Genieacs MCP Bridge to GenieACS (TR-069 ACS) instances: device data, firmware management and CPE actions such as reboot, parameter refresh and firmware d | 19 | Oct 21, 2026 | GPL-3.0 | moderate | 63 |
| Openobserve community MCP Read-only access to OpenObserve Community Edition: search logs, traces, stream schemas and dashboards without an Enterprise license. | 16 | Mar 24, 2026 | GPL-3.0 | low | 37 |
| Llmkit AI API cost tracking and budget enforcement across LLM providers: spend analytics, budget monitoring, session summaries and key management. | 15 | Oct 22, 2026 | MIT | moderate | 62 |
| MCP internet speed test Internet speed testing: download and upload speed, latency, jitter, and CDN server detection with geographic mapping. | 15 | Mar 9, 2026 | MIT | low | 37 |
| Spanlens ◆ Read-only access to Spanlens LLM observability: request logs, costs, agent traces, anomalies, model savings and per-user analytics. | 13 | Oct 19, 2026 | MIT | moderate | 71 |
| Scherlok Read-only data quality monitoring for Postgres, BigQuery, Snowflake, MySQL and DuckDB: warehouse profiling, anomaly detection and optional d | 11 | Oct 19, 2026 | MIT | moderate | 60 |
| Heimdall MCP Transparent MCP proxy that intercepts JSON-RPC messages, measures latency, stores traces in SQLite, PostgreSQL or MySQL, and exports OpenTel | 11 | Jul 16, 2026 | — | moderate | 33 |
| Edgedelta MCP server ◆ Interact with Edge Delta anomalies, query logs, patterns and events, find root causes and optimize pipelines. | 10 | Oct 20, 2026 | MIT | low | 65 |
| Iris Agent evaluation and observability: trace logging, output quality evaluation, cost tracking, built-in eval rules, a real-time dashboard and | 9 | Oct 22, 2026 | MIT | moderate | 60 |
| Webmin MCP server Webmin Linux system administration: services, users, storage, security, databases and more. | 9 | Jul 7, 2026 | MIT | low | 43 |
| MCP tautulli Read-only Tautulli (Plex monitoring) access: activity, history, library and user stats, transcode analysis and resolution breakdowns. | 9 | Jul 24, 2026 | MIT | low | 43 |
| Sentry MCP rs Minimal Sentry integration. | 9 | Mar 22, 2026 | MIT | low | 35 |
| Umami MCP Access the Umami Analytics API: websites, stats, sessions, events, reports, users, teams and realtime monitoring. | 8 | Oct 19, 2026 | MIT | moderate | 59 |
| MCP status observer Monitor the operational status of major digital platforms. | 8 | Oct 21, 2026 | MPL-2.0 | low | 54 |
| AgenticLedger Query the local Agentic Ledger flight recorder: agent sessions, costs, loop runs and stuck-loop flags captured by a transparent proxy. | 7 | Oct 24, 2026 | MIT | low | 58 |
| Llmprobe Synthetic monitoring for cloud and local LLM inference endpoints: TTFT, latency, throughput and errors, with Prometheus and OpenTelemetry ex | 7 | Oct 18, 2026 | MIT | low | 53 |
| Loki MCP server Query Grafana Loki directly with a discovery-first workflow: labels, values, series and LogQL queries, without requiring Grafana. | 7 | Oct 24, 2026 | MIT | low | 53 |
| Datadog MCP server Datadog observability: grep-like log search, APM trace filtering by duration, status and errors, sampling modes, and log, trace and metric c | 6 | Oct 24, 2026 | Apache-2.0 | moderate | 58 |
| Logmcp Read-only access to whitelisted log files on a Linux server: search and read logs without shell access, with token auth and syslog auditing. | 5 | Aug 25, 2026 | MIT | low | 52 |
| Siteaudit MCP Website audits: SEO, security headers and SSL, Lighthouse performance, broken links, WCAG accessibility, Schema.org validation, competitor g | 5 | Apr 17, 2026 | MIT | low | 37 |
| LLM Usage Local LLM API cost tracking for Anthropic, OpenAI, Qwen and DeepSeek: SQLite usage ledger, spend queries, provider comparison and CNY to USD | 5 | Oct 21, 2026 | MIT | low | 57 |
| Tallybook Local cost tracking for Claude Code and Codex sessions: spend by session, sub-agent and model, week-on-week comparisons, and findings with s | 5 | Sep 13, 2026 | MIT | low | 47 |
| MCP server logs sieve Query, summarize and trace logs in plain English across GCP Cloud Logging, AWS CloudWatch, Azure Log Analytics, Grafana Loki and Elasticsear | 5 | Oct 9, 2026 | MIT | low | 52 |
| HostTracker ◆ Website uptime monitoring: instant HTTP, ping, port, DNS, WHOIS, blacklist and page-speed checks, plus monitors, webhooks, incidents and sta | 1 | Aug 30, 2026 | MIT | low | 55 |
| Watchgoose ◆ Monitor cron jobs, backups, Kubernetes jobs and other scheduled tasks in Watchgoose. Read-only by default; check changes require explicit co | 1 | Oct 24, 2026 | MIT | moderate | 60 |
| Log10x MCP ◆ Rank log patterns by volume and cost, then compact (Splunk, self-managed Elasticsearch), tier down or offload the excess over each cap. Runs | 1 | Oct 24, 2026 | MIT | moderate | 60 |
| Uptybots MCP ◆ Uptime monitoring for websites, APIs, SSL certificates, domain expiry, ping and TCP/UDP ports: manage monitors and view incidents and uptime | 1 | Aug 21, 2026 | MIT | low | 50 |
| OpenITCOCKPIT ◆ Access openITCOCKPIT monitoring: host and service state, check history, logs, downtimes, acknowledgements, software inventory and pending se | 0 | Sep 18, 2026 | MIT | moderate | 55 |
Official servers
Published by the service's own vendor (per the awesome-mcp-servers list): they usually track the API closely and are better maintained.
- Netdata 81.1K ★
- MCP grafana 3.6K ★
- Logfire MCP 161 ★
- Metoro MCP server 51 ★
- SolidPing 26 ★
- Spanlens 13 ★
- Edgedelta MCP server 10 ★
- HostTracker 1 ★
- Watchgoose 1 ★
- Log10x MCP 1 ★
- Uptybots MCP 1 ★
- OpenITCOCKPIT 0 ★
Choosing and securing
Check compatibility with your telemetry sources and whether the server separates reading from administration. Review its maintenance, official or community ownership and query documentation. Look for usable time filters, volume controls and clear error handling. When comparing local and remote deployment, consider network access, which data leaves your environment and the license terms that apply to your intended use case.
Logs and traces may contain credentials, network addresses or personal information. Management permissions may also allow changes to alerts, incidents or public status pages. Limit accessible environments and use read access during investigations. Filter secrets before sending information to external services. Logged messages can originate from outside users, so treat their contents as evidence rather than instructions for the agent.
Before connecting a server to an agent with access to your data, run the AI agent security check.
Most recent
- Tallybook created September 10, 2026
- Watchgoose created August 26, 2026
- HostTracker created August 25, 2026
- Watch skill created July 5, 2026
- Logmcp created May 23, 2026
Other categories
Learn more
Frequently asked questions
What is the best Monitoring and observability MCP server?
On our grid (adoption, maintenance, framework), Netdata (100/100), MCP grafana (96/100), and Openstatus (91/100) come first. The right choice mostly depends on the service you need and the rights you are willing to grant.
Are these servers free?
95% declare an open-source licence and are free to use. 14 need an API key: the connected service may be paid.
Can the agent identify the cause of an incident on its own?
It can connect available evidence and suggest explanations. The quality of its analysis depends on metric coverage, logs and the context you provide. Ask for supporting evidence for each hypothesis and verify the conclusions before making changes to a production system or closing an incident investigation.
Do I need administrative permissions to read metrics?
Not always. Requirements depend on the server and the source being queried. Look for a configuration restricted to reading the relevant environments. Managing alerts or incidents may require additional permissions, which you can keep separate from investigative access and enable only for an explicitly approved operational task.
Related articles
- Build your own MCP server in 30 minutesOctober 24, 2026
- Automate your inbox with an AI agentOctober 18, 2026
- MCP, A2A, and tool calling: agent protocols explainedOctober 17, 2026
- Build an AI agent with n8n, without codingOctober 16, 2026
Neutrality. No affiliate links. No vendor pays to be reviewed, rated or ranked. Reviews are directed by François Chollet, neutral and factual, and every data point is dated and sourced. How we score