Top DevOps MCP servers

The DevOps category has the highest average scores of any category in the index — corporate-backed tools dominate the top tier. Microsoft alone has two entries in the top three. Scores above 80 are common here; this is the most mature category in the MCP ecosystem.

Kubernetes: four strong options

Kubernetes has four credible MCP servers in the top 10, which is unusual — most categories have one clear winner. Each has a different trade-off.

reza-gholizade/k8s-mcp-server (91.66)

The highest-scoring Kubernetes tool despite only 149 stars. Its score is driven by exceptional freshness and issue health signals — the maintainer is actively responding to issues and shipping commits. Written in Go. If you're looking at what the score is actually measuring, this is a good example: fewer stars but better maintenance discipline than the higher-starred options.

rohitg00/kubectl-mcp-server (88.11)

The easiest to install — npx kubectl-mcp-server or pip install kubectl-mcp-server. 849 stars, wraps kubectl and helm directly. Best for teams that want minimal setup and broad kubectl coverage without writing any configuration.

containers/kubernetes-mcp-server (84.92)

The official server from the containers organization. Unique feature: OpenShift support. If your infrastructure runs on OpenShift, this is the only server with native support. 42 contributors and consistent commits make this the safest long-term bet for enterprise use.

Flux159/mcp-server-kubernetes (80.38)

The most-starred Kubernetes server at 1,348 stars with 40 contributors. Written in TypeScript. Good community, solid maintenance, broad coverage of Kubernetes resources. Scores slightly lower than the others on freshness — last commit was March 1 — but still well-maintained.

Cloud provider coverage

Azure: two official tools

Microsoft has invested heavily in MCP tooling. azure-devops-mcp is the #1 DevOps tool overall at 97.17 — the second-highest score in the entire index. 45 contributors, actively maintained, covers pipelines, work items, repos, and boards. azure-skills sits alongside it at 92.25, providing broader Azure scenario coverage including compute, storage, and networking.

If you're running Azure infrastructure, you have the best-covered cloud environment in the MCP ecosystem.

AWS: two specialized tools

AWS has two distinct tools in the top 10 covering different needs. alexei-led/aws-mcp-server wraps the AWS CLI in a containerized environment, which is the right security posture — credentials stay in the container, not in the agent context. awslabs/Log-Analyzer-with-MCP is the official AWS Labs server, specialized for CloudWatch log analysis rather than general AWS operations.

No general-purpose "do everything in AWS" MCP server exists yet with strong quality signals. The aws-mcp-server is the best general option, but it's a community project with 2 contributors.

GCP: official coverage

Google has two official tools: cloud-run-mcp from GoogleCloudPlatform (84.69, focused on Cloud Run deployments) and gcloud-mcp from googleapis (81.99, broader gcloud CLI coverage). Both are actively maintained. Use cloud-run-mcp for deployment automation; use gcloud-mcp for general infrastructure management.

Recommended picks

Best overall DevOps tool

microsoft/azure-devops-mcp at 97.17 is the best-maintained DevOps MCP server in the index by a significant margin. If you're on Azure, this is the clear choice.

Best for Kubernetes (multi-cloud)

containers/kubernetes-mcp-server for enterprise and OpenShift environments. reza-gholizade/k8s-mcp-server if you want the highest-quality signals in a focused, fast-moving project.

Best for AWS

alexei-led/aws-mcp-server for general AWS operations. awslabs/Log-Analyzer-with-MCP specifically for log analysis and incident triage.

Best for GCP

googleapis/gcloud-mcp for broad coverage. GoogleCloudPlatform/cloud-run-mcp specifically for deployment automation.