mcp-datadog·integration·active

Datadog MCP

Streams APM metrics into LEO during load runs. ARIA correlates production incidents with sprint changes for retrospective analysis.

LEO ARIAtrigger · mcp⌘ MCP · Datadog

Datadog MCP

Streams APM metrics into LEO during load runs. ARIA correlates production incidents with sprint changes for retrospective analysis.

What flows in

·Metric time-series (latency, error rate, throughput, saturation)

·APM trace samples (for slowest-N traces during a run)

·Log streams (filtered by service + tag)

·Incident records (for ARIA's retrospective correlation)

·Dashboard definitions (so LEO can attach context to a run)

What flows out

·Run annotations on Datadog dashboards (so SREs see when LEO ran)

·Custom metrics (e.g. EVAN's pass rate published as a Datadog gauge for trend dashboards)

·Tagged events (sprint boundaries, agent run IDs) for cross-tool correlation

Tag conventions

→The connector relies on consistent tagging — service, env, version, team. If the org's Datadog isn't tagged consistently, LEO's correlation degrades.

A first-run check verifies that the target service has at least service and env tags. If not, LEO surfaces a recommendation:"tag your services consistently before relying on observability correlation".

Metric queries

The connector accepts Datadog query language directly. For LEO'sObservability Hookup, common queries are pre-baked:

·avg:postgresql.connections.active{service:auth} / avg:postgresql.connections.max{service:auth} — connection pool utilization

·p99:trace.http.request{service:auth} — endpoint p99

·sum:runtime.gc.pauses.duration{service:auth}.as_rate() — GC time

The library is small but covers the metrics most worth correlating with load tests.

Trace sampling

For each LEO run, the connector pulls the slowest 10 traces in the run window. These are attached to the run report so engineers can seewhy a request was slow, not justthat it was slow.

Incident correlation

When ARIA generates a retrospective, the connector pulls Datadog incident records from the sprint window and aligns them with code changes via Git timestamps. This is what surfaces"the auth incident on Tuesday correlates with the rate-limiting deploy on Monday".

Authentication

API key + Application key. Stored encrypted, scoped per workspace.

Rate limits

Datadog's API allows 300 req/hour for queries by default. The connector batches queries aggressively to stay under this — most LEO runs use 10–20 query calls.

Server

mcp://datadog.com/mcp

Vendor

Datadog.

MCP server endpoint

mcp://datadog.com/mcp

vendor · Datadog

Related · integration

Sprint Context Loader

sprint-context-loader

Pulls sprint metadata — name, dates, team, release tag, prior trend data — directly from connected issue trackers via MCP.

Jira MCP

mcp-jira

Bidirectional Jira connector. TARA pulls stories from sprints; ARIA pushes findings, recommendations, and quality reports back as issues, comments, or sprint reviews.

TestRail MCP

mcp-testrail

Pulls existing test cases and execution history; pushes EVAN-generated tests and run results back as cases, runs, and milestone reports.

Azure DevOps MCP

mcp-ado

Pulls work items + repos; pushes test plans, suites, and runs. Handles ADO test points and outcome states with full bi-directional sync.

GitHub MCP

mcp-github

Reads PRs, issues, code search; opens PRs containing EVAN-generated automation, SARA remediation patches, and ARIA-recommended changes.

Confluence MCP

mcp-confluence

Reads PRDs and ACs from Confluence pages as TARA input. Publishes ARIA's quality reports back as Confluence pages with embedded charts.