ClickHouse Reference Architecture

Overview

Production-grade architecture for ClickHouse analytics platforms covering project layout, data flow, multi-tenancy, and operational patterns.

Prerequisites

• Understanding of ClickHouse fundamentals (engines, ORDER BY, partitioning)
• TypeScript/Node.js project

Instructions

Step 1: Project Structure

my-analytics-platform/
├── src/
│   ├── clickhouse/
│   │   ├── client.ts           # Singleton client with health checks
│   │   ├── schemas/            # SQL DDL files (source of truth)
│   │   │   ├── 001-events.sql
│   │   │   ├── 002-users.sql
│   │   │   └── 003-materialized-views.sql
│   │   ├── queries/            # Named query functions
│   │   │   ├── events.ts
│   │   │   ├── users.ts
│   │   │   └── dashboards.ts
│   │   └── migrations/         # Schema migrations
│   │       ├── runner.ts
│   │       └── 001-add-country.sql
│   ├── ingestion/
│   │   ├── webhook-receiver.ts # HTTP webhook endpoint
│   │   ├── kafka-consumer.ts   # Kafka consumer (if applicable)
│   │   └── buffer.ts           # Insert batching buffer
│   ├── api/
│   │   ├── routes.ts           # API endpoints
│   │   └── middleware.ts       # Auth, rate limiting
│   └── jobs/
│       ├── daily-rollup.ts     # Scheduled aggregations
│       └── cleanup.ts          # TTL enforcement
├── tests/
│   ├── unit/
│   └── integration/
├── docker-compose.yml          # Local ClickHouse
├── init-db/                    # Docker init scripts
└── config/
    ├── development.env
    ├── staging.env
    └── production.env

Step 2: Data Flow Architecture

                    ┌─────────────────┐
                    │   Data Sources   │
                    │  (Webhooks, API, │
                    │   Kafka, S3)     │
                    └────────┬────────┘
                             │
                    ┌────────▼────────┐
                    │  Ingestion Layer │
                    │  (Buffer + Batch │
                    │   10K+ rows/ins) │
                    └────────┬────────┘
                             │
              ┌──────────────▼──────────────┐
              │       ClickHouse Server      │
              │                              │
              │  ┌────────────────────────┐  │
              │  │   Raw Event Tables     │  │
              │  │   (MergeTree, append)  │  │
              │  └───────────┬────────────┘  │
              │              │               │
              │  ┌───────────▼────────────┐  │
              │  │  Materialized Views    │  │
              │  │  (Auto-aggregate on    │  │
              │  │   INSERT — hourly,     │  │
              │  │   daily, tenant-level) │  │
              │  └───────────┬────────────┘  │
              │              │               │
              │  ┌───────────▼────────────┐  │
              │  │  Aggregate Tables      │  │
              │  │  (AggregatingMergeTree)│  │
              │  └────────────────────────┘  │
              └──────────────┬──────────────┘
                             │
                    ┌────────▼────────┐
                    │    API Layer     │
                    │  (Query aggregate│
                    │   tables, not    │
                    │   raw events)    │
                    └────────┬────────┘
                             │
                    ┌────────▼────────┐
                    │   Dashboards /   │
                    │   Client Apps    │
                    └─────────────────┘

Step 3: Schema Design (3-Layer Pattern)

-- Layer 1: Raw events (append-only, full fidelity)
CREATE TABLE analytics.events_raw (
    event_id    UUID DEFAULT generateUUIDv4(),
    tenant_id   UInt32,
    event_type  LowCardinality(String),
    user_id     UInt64,
    properties  String CODEC(ZSTD(3)),
    created_at  DateTime64(3) DEFAULT now64(3)
)
ENGINE = MergeTree()
ORDER BY (tenant_id, event_type, toDate(created_at), user_id)
PARTITION BY toYYYYMM(created_at)
TTL created_at + INTERVAL 90 DAY;

-- Layer 2: Hourly aggregation (auto-populated via materialized view)
CREATE TABLE analytics.events_hourly (
    hour        DateTime,
    tenant_id   UInt32,
    event_type  LowCardinality(String),
    cnt         UInt64,
    users       AggregateFunction(uniq, UInt64)
)
ENGINE = AggregatingMergeTree()
ORDER BY (tenant_id, event_type, hour);

CREATE MATERIALIZED VIEW analytics.events_hourly_mv TO analytics.events_hourly AS
SELECT toStartOfHour(created_at) AS hour, tenant_id, event_type,
       count() AS cnt, uniqState(user_id) AS users
FROM analytics.events_raw GROUP BY hour, tenant_id, event_type;

-- Layer 3: Daily rollup for dashboards
CREATE TABLE analytics.events_daily (
    date        Date,
    tenant_id   UInt32,
    total       UInt64,
    users       AggregateFunction(uniq, UInt64)
)
ENGINE = AggregatingMergeTree()
ORDER BY (tenant_id, date);

CREATE MATERIALIZED VIEW analytics.events_daily_mv TO analytics.events_daily AS
SELECT toDate(created_at) AS date, tenant_id,
       count() AS total, uniqState(user_id) AS users
FROM analytics.events_raw GROUP BY date, tenant_id;

Step 4: Multi-Tenant Patterns

Approach A: Shared table with tenant_id in ORDER BY (recommended)

-- Tenant_id first in ORDER BY = queries filter on tenant efficiently
ORDER BY (tenant_id, event_type, created_at)

-- Query: only scans data for this tenant
SELECT count() FROM events_raw WHERE tenant_id = 42;

Approach B: Database per tenant (for strict isolation)

CREATE DATABASE tenant_42;
CREATE TABLE tenant_42.events (...) ENGINE = MergeTree() ...;

-- Pros: Full isolation, easy to drop tenant
-- Cons: Schema changes need per-tenant DDL, more operational overhead

Approach C: Row-level security (ClickHouse RBAC)

CREATE ROW POLICY tenant_isolation ON analytics.events_raw
    FOR SELECT USING tenant_id = getSetting('custom_tenant_id')
    TO app_user;

Step 5: Client Module

// src/clickhouse/client.ts
import { createClient, ClickHouseClient } from '@clickhouse/client';

let instance: ClickHouseClient | null = null;

export function getClient(): ClickHouseClient {
  if (!instance) {
    instance = createClient({
      url: process.env.CLICKHOUSE_HOST!,
      username: process.env.CLICKHOUSE_USER!,
      password: process.env.CLICKHOUSE_PASSWORD!,
      database: process.env.CLICKHOUSE_DATABASE ?? 'analytics',
      max_open_connections: Number(process.env.CH_MAX_CONNECTIONS ?? 10),
      request_timeout: 30_000,
      compression: { request: true, response: true },
    });
  }
  return instance;
}

// src/clickhouse/queries/dashboards.ts
export async function getTenantDashboard(tenantId: number, days = 30) {
  const client = getClient();
  const rs = await client.query({
    query: `
      SELECT date, sum(total) AS events, uniqMerge(users) AS unique_users
      FROM analytics.events_daily
      WHERE tenant_id = {tid:UInt32} AND date >= today() - {days:UInt32}
      GROUP BY date ORDER BY date
    `,
    query_params: { tid: tenantId, days },
    format: 'JSONEachRow',
  });
  return rs.json<{ date: string; events: string; unique_users: string }>();
}

Architecture Decision Records

Decision	Choice	Why
Engine	MergeTree (raw) + AggregatingMergeTree (rollups)	Best for append + pre-agg
Multi-tenant	Shared table + tenant_id in ORDER BY	Scales to 10K+ tenants
Ingestion	Buffer + batch INSERT	Avoids "too many parts"
Aggregation	Materialized views (not cron)	Real-time, zero-lag
Format	JSONEachRow	Client support, debugging
Compression	ZSTD(3) for strings, Delta for ints	10-20x compression

Error Handling

Issue	Cause	Solution
Cross-tenant data leak	Missing WHERE tenant_id	Use row policies or middleware
Stale dashboard data	MV not created	Verify MV exists and is attached
Schema drift	Manual DDL changes	Use migration runner
Slow dashboard queries	Querying raw table	Query aggregate tables instead

Resources

• ClickHouse Architecture
• SharedMergeTree (Cloud)
• Materialized Views

Next Steps

For multi-environment configuration, see clickhouse-multi-env-setup.