PROVIDER Β· High-performance time-series database

QuestDB

QuestDB is a time-series database built for high ingestion throughput, where columnar storage and time partitioning with SAMPLE BY and LATEST ON fit market data, IoT, and metric workloads.

Backend & Database United Kingdom 2026.11.15
Backend & Database United Kingdom High-performance time-series database
3.9pts

Pros

4 items
Columnar time-series storage with vendor-claimed ingestion near a million rows per second
Built-in SAMPLE BY and LATEST ON make downsampling and latest-value queries intuitive
InfluxDB line protocol and PostgreSQL wire protocol lower integration cost
Time partitioning with vectorised execution keeps wide time aggregations steady

Cons

4 items
Time-series only, lacking general transactions and complex joins
The open source edition is mostly single-node with clustering in commercial versions
Smaller community and ecosystem than InfluxDB or TimescaleDB
Scarce Chinese material, requiring self-accumulated operations experience

Product Information

questdb
Category Backend & Database
Company QuestDB
Service Region United Kingdom
Official Website https://questdb.com/
Recommendation Score 3.9 pts
Published 2026-11-15

Detailed Notes

Overview

QuestDB was open sourced in 2019 by a UK team and written in Java and C++, purpose-built for time series with columnar storage, time partitioning, and vectorised execution, claiming millions of rows per second ingestion alongside standard SQL and InfluxDB line protocol compatibility.

For Backend & Database work the advantage is writes that keep up with reads, since market data, sensors, and metrics ingest without write amplification through write-ahead logging and columnar appends, and built-in SAMPLE BY and LATEST ON make downsampling and latest-value lookups far clearer than generic SQL.

The limits are narrow scope and a small ecosystem, because it suits time series only without general transactions or complex joins, community and third-party tooling trail InfluxDB and TimescaleDB, and clustering plus high availability are commercial with the open source edition mostly single-node.

Core Strengths

  • Columnar time-series storage with vendor-claimed ingestion near a million rows per second
  • Built-in SAMPLE BY and LATEST ON make downsampling and latest-value queries intuitive
  • InfluxDB line protocol and PostgreSQL wire protocol lower integration cost
  • Time partitioning with vectorised execution keeps wide time aggregations steady

Product Ecosystem

Time-series tables and partitioning

Storage partitions by time with configurable granularity and TTL cleanup.

Ingestion protocols

InfluxDB line protocol, PostgreSQL wire protocol, and REST API fit many collectors.

Deployment

Open source single node, Docker or Kubernetes, and QuestDB Cloud managed hosting.

Drawbacks

  • Time-series only, lacking general transactions and complex joins
  • The open source edition is mostly single-node with clustering in commercial versions
  • Smaller community and ecosystem than InfluxDB or TimescaleDB
  • Scarce Chinese material, requiring self-accumulated operations experience

Use Cases

  • Financial market and tick data (β˜…β˜…β˜…β˜…β˜…) - High ingestion with low-latency queries
  • IoT sensor data (β˜…β˜…β˜…β˜…) - Columnar compression with TTL
  • Metrics and log analysis (β˜…β˜…β˜…β˜…) - Time-series SQL syntax
  • General business data (β˜…) - Use a relational database

Pricing

Plan Price Scope
QuestDB open source $0 Apache 2.0, single-node self-managed
QuestDB Cloud Per instance and usage Managed with backup and scaling
Enterprise Per node subscription Clustering and high availability

The open source edition is free and a single node ingests very high volume with cost in servers and disks, while multi-node HA enters commercial plans, and storage cost tracks retention policy with compression usually beating general databases.

FAQ

  • QuestDB or InfluxDB? QuestDB for standard SQL, high ingestion throughput, and simple single-node deployment, InfluxDB or TimescaleDB for mature ecosystems, clustering, and full TSDB features.

  • How much data fits? A single node holds terabytes with sensible partitioning and TTL, while longer history should archive to object storage by time partition.

  • Are concurrent writes supported? Multiple clients can append concurrently though per-table writes still depend on disk and partitioning, so very high concurrency benefits from table splits or a buffer queue.

Visit QuestDB Website β†’ https://questdb.com/
Visit Website β†—