How to build LaunchDarkly dashboards in Metabase

LaunchDarkly is a feature management platform for feature flags, progressive rollouts, and experimentation at enterprise scale. Metabase is where you turn that delivery activity into shared, trustworthy dashboards. This guide covers two complementary paths: a lightweight MCP + CLI route that pulls live data with the LaunchDarkly MCP Server and loads a CSV into Metabase with the Metabase CLI, and a durable pipeline route that syncs LaunchDarkly history into a database so you can build dashboards anyone can read.

How do you connect LaunchDarkly to Metabase?

Most teams combine both routes: use MCP and CLI uploads for a fast first pass, then move recurring delivery reporting to a warehouse-backed model.

What can you analyze from LaunchDarkly data in Metabase?

Once feature flags and the related flag changes (audit log), environments, experiments data are in a database, the highest-value views are:

• Flag rollout activity — changes per environment per week, straight from audit-log events.
• Stale flag cleanup — temporary flags past 90 days — the debt list to burn down.
• Change activity vs. incidents — flag changes on the same timeline as incidents and error spikes.
• Experiment outcomes — what launched, what concluded, and what actually won.
• Flag debt by project — where flags accumulate fastest, and which teams own the cleanup.

Which LaunchDarkly dashboards should you build in Metabase?

How do you use the LaunchDarkly MCP Server with the Metabase CLI?

Pair the LaunchDarkly MCP Server with the Metabase CLI for fast, hands-on analysis. MCP is useful for scoped lookups and summarized exports; the Metabase CLI's upload command loads CSV data into Metabase and creates a ready-to-query table and model.

Example workflow

• Ask the MCP server for the current feature flags inventory with recent change history and rollout status.
• Export the result as CSV, keeping stable IDs, projects, states, and timestamps.
• Run mb upload csv to load it into Metabase as a table and model, then build questions and dashboards on top.

Be honest about the limits

• MCP lookups are excellent for exploration, not scheduled reporting.
• A CSV upload is a snapshot; refresh it with mb upload replace or move to the pipeline for real history.
• Audit-log change events are required for rollout velocity and change-vs-incident analysis.
• mb upload csv needs an uploads database configured under Admin → Settings → Uploads.

How do you set up LaunchDarkly MCP and the Metabase CLI?

{
  "mcpServers": {
    "launchdarkly": {
      "url": "https://mcp.launchdarkly.com/mcp/launchdarkly"
    }
  }
}

EU and federal instances should run the official local package with an API access token instead of the hosted endpoint.

# Install the Metabase CLI
npm install -g @metabase/cli

# Log in (opens your browser; requires Metabase v62+)
mb auth login --url https://your-metabase.example.com

# Load a feature-flags export — creates a table AND a model
mb upload csv --file launchdarkly-feature-flags.csv --collection root

# Refresh that same table later from a new export
mb upload replace <table-id> --file launchdarkly-feature-flags.csv

Can you generate a LaunchDarkly dashboard with AI?

Yes. Use the prompt below with any assistant that can run the LaunchDarkly MCP Server and the Metabase CLI. It works end to end: if LaunchDarkly tables already exist in Metabase it analyzes those; otherwise it pulls scoped, summarized data over MCP, loads it with mb upload csv, then builds the dashboard and caveats any metric that needs missing history.

Create a polished Metabase dashboard for LaunchDarkly feature flag analytics.
Work end to end: get the data into Metabase if it isn't there yet, then build.

Goal: Help engineering leaders understand rollout activity, flag hygiene, experiment outcomes, and change safety from LaunchDarkly data.

Step 1 — Find or load the data:
- First, check what already exists in Metabase (search for launchdarkly tables and
  models). If durable LaunchDarkly data is already present — synced from a warehouse
  or uploaded earlier — use it and skip to Step 2.
- If nothing is there, pull a scoped, summarized export with the LaunchDarkly MCP Server:
  feature flags, plus flag changes (audit log), environments, experiments.
  Prefer entity tables and rollups over raw logs. Write each result to a CSV,
  then load it with the Metabase CLI — run "mb upload csv --file <export>.csv" so
  each upload creates a table and a ready-to-query model. Use "mb upload replace
  <table-id> --file <export>.csv" to refresh an existing table instead of creating
  duplicates.

Step 2 — Inspect before querying:
Do not assume exact table or column names. Inspect available fields, projects,
branches, states, and timestamps, and whether history exists before creating
duration or trend cards.

Important:
- Build on whatever data is present; don't claim Metabase connects natively to
  LaunchDarkly — it reads a database or CLI-uploaded tables.
- Never try to load raw build logs or event streams into Metabase; use runs,
  entities, state changes, and rollups.
- Only compute durations (cycle time, build duration, time to review) when the
  required timestamps exist.
- Exclude bot activity, draft work, and non-production environments from
  headline numbers, and keep the segmenting column explicit.
- A single CSV is a point-in-time snapshot: only build trend cards if there is a
  usable date column or multiple periods have been uploaded.

Dashboard title: LaunchDarkly Feature Flag Overview

Sections:
1. Executive summary: Active flags; Changes last 30 days; Temporary flags
   over 90 days old; Experiments running; Experiments concluded last quarter.
2. Rollouts: Changes per week by environment; in-progress rollouts; time to
   full rollout.
3. Hygiene: Stale temporary flags; unevaluated flags; flags per project.
4. Experiments: Launched vs concluded; win/loss share; days to decision.
5. Change safety: Changes vs incidents timeline; changes followed by error
   spikes; off-hours changes.

Filters: Date range, Project or Repository, Branch or Environment, Team, Status.

Output: Build the dashboard if you have permission; otherwise provide the exact
questions, SQL, model definitions, and layout. Include caveats for any metric
that cannot be calculated from the available data.

How do you sync LaunchDarkly data into a database or warehouse?

For dashboards that need history and reliability, land LaunchDarkly entities and change history in a database first, then connect Metabase to that database.

Connector options

• Managed ETL — use a connector when one covers the objects you need.
• Custom pipeline — use the LaunchDarkly REST API for control over fields, grain, and refresh cadence.
• MCP + CSV — use this for quick exploration and one-off slices.

Fivetran ships an official LaunchDarkly connector and Airbyte has a community source — both cover flags, environments, projects, and audit-log entries. The audit log is the analytics gold: land it as a change-events table and you can join flag changes to deploys and incidents.

Notes

• Land raw entity tables first, then build clean Metabase models on top.
• Sync state-change or audit history from day one — duration metrics can't be reconstructed later without it.
• Normalize project, environment, flag key, action, actor, and changed-at fields.

How should you model LaunchDarkly data in Metabase?

Core tables

Modeling advice

• Build a clean flag_changes model with common columns across tools, so multi-source dashboards don't fork definitions.
• Separate entity tables (projects, repos, flags, members) from run-grain and event-grain tables.
• Exclude bots, drafts, and non-production environments from headline metrics; keep the flag as an explicit column.
• Use stable IDs for project, repo, and person joins; display names change.

Which LaunchDarkly metrics should you track in Metabase?

What SQL powers LaunchDarkly dashboards in Metabase?

These assume a cleaned analytical model in a warehouse (PostgreSQL dialect). Adjust table and column names to match your pipeline.

SELECT
  date_trunc('week', changed_at) AS week,
  environment,
  COUNT(*) AS flag_changes
FROM flag_changes
WHERE changed_at >= CURRENT_DATE - INTERVAL '90 days'
GROUP BY 1, 2
ORDER BY 1, 2;

SELECT
  key,
  project,
  created_at,
  EXTRACT(DAY FROM (CURRENT_TIMESTAMP - created_at)) AS age_days
FROM flags
WHERE temporary
  AND NOT archived
  AND created_at < CURRENT_DATE - INTERVAL '90 days'
ORDER BY created_at
LIMIT 50;

SELECT
  date_trunc('month', decision_made_at) AS month,
  COUNT(*) AS experiments_concluded,
  COUNT(*) FILTER (WHERE winning_variant IS NOT NULL) AS with_winner,
  percentile_cont(0.5) WITHIN GROUP (
    ORDER BY EXTRACT(EPOCH FROM (decision_made_at - started_at)) / 86400
  ) AS median_days_to_decision
FROM experiments
WHERE decision_made_at IS NOT NULL
GROUP BY 1
ORDER BY 1;

What are common mistakes when analyzing LaunchDarkly in Metabase?

Related

Related analytics

Related dashboards

Related integrations

FAQ