$ loading_
Securely query metrics, alerts, and logs for read-only cluster diagnostics.
This tool is described as an in-cluster, read-only diagnostic MCP server with no extra secrets and no declared external remote endpoints; based on the stated functionality, the main risks come from running a local service and accessing in-cluster observability data. However, only a brief description is available, the README is missing, adoption is minimal, and maintenance is unknown, so its supply-chain trust and actual permission boundaries should be verified carefully.
The materials explicitly state that no secrets or environment variables are required; it claims to use Kubernetes RBAC authentication rather than additional tokens, with no visible requirement for users to provide new sensitive credentials or obvious credential-abuse design.
Although no external remote host is declared, the description indicates it communicates in-cluster with Grafana, Prometheus, Alertmanager, and Loki via Grafana datasource proxies; this is network access consistent with its stated purpose, potentially carrying metrics, alerts, and log queries, so users should verify it is limited to expected in-cluster endpoints.
The system checks indicate the tool executes code / runs a service process; for an MCP tool this is a normal capability, and the materials do not show requests for extra system execution privileges unrelated to diagnostics, nor suspicious command-execution behavior.
By description, it can read observability data including metrics, alert listings, and log query results; while it is presented as read-only diagnostics, such data may still contain business metadata or sensitive log fragments, so its RBAC scope should be verified as minimal and truly read-only.
A positive factor is that there is an open-source repository, which provides some auditability; however, it comes from a third-party registry, has no README, no declared license, 0 stars, and unknown maintenance status, so trust is limited and the source code and dependency tree should be reviewed before use.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "mcp-grafana" yet — see the docs or source repo.
Using mcp-grafana, query the last 30 minutes of P95 latency, error rate, and related logs for payment-service, then summarize possible causes.
Returns relevant metric query results, matching log excerpts, and a brief analysis of likely latency causes.
List all currently firing Grafana or Alertmanager alerts in the production cluster, sorted by severity, and label the owning service.
Returns a severity-sorted alert list including alert name, status, service, and key labels.
Use mcp-grafana to find the most frequently restarting pods in the orders namespace over the last 24 hours, and use Prometheus metrics plus Loki logs to determine whether OOM or probe failures are involved.
Returns the top restarting pods, relevant resource and status metrics, key logs, and an initial fault assessment.
Connect to Grafana dashboards, data sources, and alerts for monitoring analysis.
Access Grafana dashboards and data to inspect metrics and visual insights.
Access Grafana dashboards, datasources, and metric or log queries read-only.
Manage Kubernetes clusters, resources, backups, and diagnostics using natural language.
Query Prometheus metrics with PromQL and analyze monitoring trends and anomalies.
Connect to Kubernetes via MCP to inspect clusters and assist operations.