How to build RudderStack dashboards in Metabase

RudderStack is a warehouse-first customer data platform that routes events into your own warehouse and on to downstream tools. Metabase is where you turn that marketing data into shared, trustworthy dashboards. This guide covers two complementary paths: a lightweight MCP + CLI route that pulls live data with the RudderStack MCP server and loads a CSV into Metabase with the Metabase CLI, and a durable pipeline route that syncs RudderStack daily stats into a database so you can build dashboards anyone can read.

How do you connect RudderStack to Metabase?

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

What can you analyze from RudderStack data in Metabase?

• Event volume by source over time — built from event volume by source and the related sources, destinations, warehouse sync runs data your sync exposes.
• Delivery success by destination — built from event volume by source and the related sources, destinations, warehouse sync runs data your sync exposes.
• Warehouse sync latency and failures — built from event volume by source and the related sources, destinations, warehouse sync runs data your sync exposes.
• Transformation error rates — built from event volume by source and the related sources, destinations, warehouse sync runs data your sync exposes.
• Identified vs. anonymous profile share — built from event volume by source and the related sources, destinations, warehouse sync runs data your sync exposes.

Which RudderStack dashboards should you build in Metabase?

How do you use the RudderStack MCP server with the Metabase CLI?

Pair the RudderStack 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 last 30 days of event volume by source by source and destination, with statuses.
• Export the result as CSV, keeping stable IDs, channels, campaigns, and dates.
• 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.
• Per-run sync outcomes and daily event counts are required for reliability and volume trends.
• mb upload csv needs an uploads database configured under Admin → Settings → Uploads.

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

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

The server answers pipeline questions in natural language — event flow, failing destinations, transformation behavior. Writes are deliberately narrow: it can create and connect transformations, but not create or delete sources and destinations.

# 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 event-volume-by-source export — creates a table AND a model
mb upload csv --file rudderstack-event-volume-by-source.csv --collection root

# Refresh that same table later from a new export
mb upload replace <table-id> --file rudderstack-event-volume-by-source.csv

Can you generate a RudderStack dashboard with AI?

Yes. Use the prompt below with any assistant that can run the RudderStack MCP server and the Metabase CLI. It works end to end: if RudderStack 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 RudderStack customer data analytics.
Work end to end: get the data into Metabase if it isn't there yet, then build.

Goal: Help marketing and growth leaders understand event volume, sync reliability, identity resolution, and whether downstream tools receive fresh data from RudderStack data.

Step 1 — Find or load the data:
- First, check what already exists in Metabase (search for rudderstack tables and
  models). If durable RudderStack 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 RudderStack MCP server:
  event volume by source, plus sources, destinations, warehouse sync runs.
  Prefer daily aggregates over raw events. 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, channels,
campaigns, dates, and whether daily history exists before creating trend or
pacing cards.

Important:
- Build on whatever data is present; don't claim Metabase connects natively to
  RudderStack — it reads a database or CLI-uploaded tables.
- Never try to load raw event or click streams into Metabase; use daily
  aggregates, campaign-grain stats, and entity tables.
- Only compute rates (CTR, conversion rate, ROAS, CAC) when both numerator and
  denominator exist — and state the attribution model when reporting conversions.
- Exclude test campaigns and internal traffic from headline cards, and keep
  currency consistent when spend spans accounts.
- 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: RudderStack Customer Data Overview

Sections:
1. Executive summary: Events last 7 days; Sync success rate; Stale
   destinations; Identified profile share.
2. Pipeline: Sync success rate and failed runs by destination; latency trend.
3. Volume: Events by source by day; anomalies vs. trailing average.
4. Quality: Schema violations, blocked events, and new event names.
5. Audiences: Sizes over time; sync freshness; rows delivered to activation tools.

Filters: Date range, Channel, Campaign, Country, Device, Segment.

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 RudderStack data into a database or warehouse?

For dashboards that need history and reliability, land RudderStack daily stats and entities 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 RudderStack warehouse destinations for control over grain, fields, and refresh cadence.
• MCP + CSV — use this for quick exploration and one-off slices.

RudderStack's warehouse-first design means the pipeline is built in: point a warehouse destination at Snowflake, BigQuery, Postgres, or a dozen others and events, identifies, and per-event tables land on your sync schedule. No third-party connector needed.

Notes

• Decide the rollup grain first (daily per campaign/channel is the workhorse) — it drives warehouse cost and every trend card.
• Land raw entity tables first, then build clean Metabase models on top.
• Normalize source, destination, event name, date, event counts, sync status, rows moved, and error counts fields.

How should you model RudderStack data in Metabase?

Core tables

Modeling advice

• Build a clean event_volume_daily model with common columns across tools, so multi-channel dashboards don't fork definitions.
• Separate entity tables (campaigns, audiences, pages) from daily time-series rollups.
• Exclude test campaigns and internal traffic from headline metrics; keep channel and campaign as explicit columns.
• Use stable IDs for campaign, channel, and user joins; display names change.

Which RudderStack metrics should you track in Metabase?

What SQL powers RudderStack dashboards in Metabase?

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

SELECT
  destination,
  date_trunc('week', run_started_at) AS week,
  COUNT(*) AS runs,
  COUNT(*) FILTER (WHERE status = 'success') AS successful_runs,
  ROUND(
    100.0 * COUNT(*) FILTER (WHERE status = 'success')
    / NULLIF(COUNT(*), 0), 2
  ) AS success_rate_pct,
  SUM(rows_synced) AS rows_synced
FROM cdp_sync_runs
GROUP BY 1, 2
ORDER BY 1, 2;

WITH daily AS (
  SELECT source, event_date, SUM(event_count) AS events
  FROM event_volume_daily
  GROUP BY 1, 2
)
SELECT
  source,
  event_date,
  events,
  ROUND(AVG(events) OVER (
    PARTITION BY source
    ORDER BY event_date
    ROWS BETWEEN 28 PRECEDING AND 1 PRECEDING
  )) AS trailing_28d_avg,
  ROUND(100.0 * events / NULLIF(AVG(events) OVER (
    PARTITION BY source
    ORDER BY event_date
    ROWS BETWEEN 28 PRECEDING AND 1 PRECEDING
  ), 0) - 100, 1) AS pct_vs_trailing
FROM daily
ORDER BY source, event_date DESC;

SELECT
  destination,
  MAX(run_finished_at) FILTER (WHERE status = 'success' AND COALESCE(rows_synced, 0) > 0)
    AS last_successful_sync,
  ROUND(EXTRACT(EPOCH FROM (
    NOW() - MAX(run_finished_at) FILTER (WHERE status = 'success' AND COALESCE(rows_synced, 0) > 0)
  )) / 3600, 1) AS hours_since_success
FROM cdp_sync_runs
GROUP BY destination
HAVING MAX(run_finished_at) FILTER (WHERE status = 'success' AND COALESCE(rows_synced, 0) > 0)
  < NOW() - INTERVAL '24 hours'
ORDER BY hours_since_success DESC;

What are common mistakes when analyzing RudderStack in Metabase?

Related

Related analytics

Related dashboards

Related integrations

FAQ