# edgedelta/edgedelta-mcp-server [Health: Active]

**Category:** 📊 Monitoring  
**Repository:** https://github.com/edgedelta/edgedelta-mcp-server  
**GitHub Stars:** 9  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/edgedelta-edgedelta-mcp-server

## Description
Interact with Edge Delta anomalies, query logs / patterns / events, and pinpoint root causes and optimize your pipelines.

## Claude Desktop Quick Installation
Heuristic fallback — verify the package name and runner against the repository README before running it. Uses `npx` (confidence: low):

```json
"mcpServers": {
  "edgedelta-mcp-server": {
    "command": "npx",
    "args": ["-y","edgedelta-edgedelta-mcp-server"],
    "env": {
      "ED_API_TOKEN": "",
      "ED_ORG_ID": ""
    }
  }
}
```

**Requires environment variables:** `ED_API_TOKEN`, `ED_ORG_ID` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What edgedelta/edgedelta-mcp-server MCP server does

The edgedelta/edgedelta-mcp-server MCP server exposes Edge Delta capabilities through the Model Context Protocol. It supports observability workflows involving anomalies, logs, patterns, and events, along with inspection and modification of Edge Delta pipelines.

Pipeline tools cover discovery, configuration, version history, validation, saving, deployment, rollback, and adding data-source nodes. Configuration changes use a controlled sequence: validate the YAML, save it as a version, and deploy that saved version. A saved version is not deployed automatically, and earlier versions can be restored with the revert operation.

The server also supports two forms of processor verification. Offline tests apply OTTL statements or complete processor nodes to sample OpenTelemetry log items without saving configuration, deploying a pipeline, or using live traffic. Live captures sample data from running agents so an agent can compare the input and output around pipeline nodes.

## How it works

An agent can first retrieve a pipeline configuration, modify the YAML, and submit it for validation. If validation succeeds, it can save the configuration, inspect the resulting history, and deploy the selected version. Revert is available when a previously saved version needs to be restored.

For live verification, the agent starts a capture task with a duration and optional node selection or item limit. It polls the task status until an agent reports success, then retrieves before-and-after samples for comparison. This distinguishes a syntactically valid configuration from a transformation that actually changed flowing data.

Offline processor tests accept sample log items. `test_pipeline_ottl` evaluates OTTL statements, including pattern-based field extraction, while `test_pipeline_node` tests processor nodes such as log-to-metric, log-to-pattern, metric extraction, JSON attribute parsing, grok, and masking.

## Setup and configuration

Running the edgedelta/edgedelta-mcp-server MCP server requires Docker Engine 20.10 or newer with Docker Buildx available. Docker Desktop includes Buildx on macOS and Windows; Linux users may need to install the `docker-buildx-plugin` package or use the documented manual fallback.

The server requires an Edge Delta API token with the necessary scope and an Edge Delta organisation ID. Configure these values through `ED_API_TOKEN` and `ED_ORG_ID`. The documented Cursor setup runs the container interactively, removes it after exit, passes both environment variables, and uses the image `ghcr.io/edgedelta/edgedelta-mcp-server:latest`.

The repository also documents a local image workflow: create and bootstrap a Buildx builder, then build the project image with the tag `mcp/edgedelta` and load it into the local Docker daemon.

## Tools and capabilities

- Discover pipelines and read configurations and version history.
- Validate YAML without saving it.
- Save, deploy, and revert pipeline versions.
- Add a data-source node to a pipeline.
- Start, monitor, inspect, and retrieve live pipeline captures.
- Dry-run OTTL transformations against sample OTEL log items.
- Dry-run complete processor nodes and inspect emitted items by output path.
- Query Edge Delta observability data involving anomalies, logs, patterns, and events.

## Limitations and notes

Saving a configuration does not deploy it. Configuration validation confirms that the configuration is valid, but live capture is needed to check the resulting data transformation on running agents. Offline processor tests do not use live traffic and do not change pipeline state.

The exported Go library API is experimental and may change without notice. The provided material documents Cursor usage; compatibility with other MCP clients is not established here. Access also depends on a correctly scoped Edge Delta API token and the organisation ID.

_Full upstream README: https://allmcps.com/mcp/edgedelta-edgedelta-mcp-server/readme_

