Hanzo

Hanzo Datastore

Managed column-oriented analytics store — run sub-second SQL over billions of rows with a ClickHouse-compatible OLAP engine.

Hanzo Datastore

Hanzo Datastore is a managed, column-oriented analytics database for OLAP workloads. It runs sub-second aggregations over billions of rows using a ClickHouse-compatible engine, exposed over SQL, an HTTP interface, and the native binary protocol.

Provision a Datastore

Each datastore is a logical resource scoped to your organization and provisioned through the control plane.

curl -X POST https://api.hanzo.ai/v1/datastore \
  -H "Authorization: Bearer hk-..." \
  -H "X-Org-Id: org_a1b2c3" \
  -H "Content-Type: application/json" \
  -d '{ "name": "events" }'

The response returns the SQL/HTTP endpoint and credentials.

Query

Datastore speaks SQL over HTTP (port 8123) and the ClickHouse native protocol (port 9000), so ClickHouse clients and drivers connect directly.

# Create a table (MergeTree family)
curl https://api.hanzo.ai/v1/datastore/events/query \
  -H "Authorization: Bearer hk-..." \
  -d 'CREATE TABLE hits (
        ts DateTime, user_id UInt64, path String, ms UInt32
      ) ENGINE = MergeTree ORDER BY (ts, user_id)'

# Aggregate
curl https://api.hanzo.ai/v1/datastore/events/query \
  -H "Authorization: Bearer hk-..." \
  -d "SELECT path, count() AS hits, avg(ms) AS mean_ms
      FROM hits
      WHERE ts > now() - INTERVAL 1 DAY
      GROUP BY path ORDER BY hits DESC LIMIT 20"

Features

  • MergeTree engine family with primary-key ordering, partitioning, and TTL
  • Real-time inserts with background merges — no batch load windows
  • Materialized views for incremental roll-ups and pre-aggregation
  • Columnar compression that routinely reaches 10x or better
  • Distributed queries and replication for horizontal scale
  • Reads cold data directly from Storage (S3-compatible buckets)

Datastore is built for append-heavy analytics — events, logs, metrics, and time series — not transactional workloads. For OLTP, use SQL.

How is this guide?

Last updated on

On this page