mcp-server-google-cloud-logging

What is mcp-server-google-cloud-logging?

An MCP server that retrieves logs from Google Cloud Logging using the gcloud CLI. It is designed for AI agents and developers who need to query logs efficiently, reducing data volume by focusing on essential fields. Works with any Google Cloud organization or project.

How to use mcp-server-google-cloud-logging?

No explicit installation or configuration steps are provided in the README. The server wraps the gcloud logging read command and exposes a tool that accepts a log query. Users must have gcloud installed and authenticated. Details on invoking the tool or setting up the server are not given.

Key features of mcp-server-google-cloud-logging

• Retrieves logs from Google Cloud Logging via gcloud.
• Returns essential fields: jsonPayload and timestamp by default.
• Optionally includes metadata fields: labels and resource.
• Reduces data volume compared to raw log output.
• Works across organizations and projects.
• Designed for accurate querying by AI agents.

Use cases of mcp-server-google-cloud-logging

• AI agents debugging application issues by querying recent logs.
• Monitoring specific services by filtering with custom log queries.
• Collecting structured log payloads for automated analysis.

FAQ from mcp-server-google-cloud-logging

What fields are included in the log output?

By default, the server returns jsonPayload (main information) and timestamp (log generation time). labels and resource (including project_id) are optional and can be included via a show option.

Does the server work across different Google Cloud organizations or projects?

Yes, it works with any organization or project that the authenticated gcloud account has access to.

What is the underlying command used?

The server uses the gcloud logging read command from the Google Cloud SDK.

What transport or authentication is required?

The README does not specify transport protocol or authentication details. The server relies on gcloud’s existing authentication, so valid credentials must be configured.

Are there any known limits?

No explicit limits are mentioned. The server aims to reduce data volume, but query accuracy depends on how logs are structured in the target project.