QuestDB
QuestDB is an open-source time-series database management system engineered for ultra-low-latency and high-throughput workloads.
Publisher review
QuestDB is an open-source time-series database management system engineered for ultra-low-latency and high-throughput workloads. It is purpose-built for latency-sensitive environments such as trading floors, mission control centers, and real-time analytics in aerospace, defense, energy, and robotics. The database is designed to ingest streaming data at millions of rows per second and query it with standard SQL, making it accessible to AI agents and developers without requiring specialized query languages. QuestDB is written in zero-GC Java and C++ to minimize garbage collection pauses, and its architecture supports a multi-tier storage engine that automatically tiers hot data in memory and cold data to Parquet files in object storage, enabling a unified SQL query surface across both tiers. The platform is deployed at tier-1 firms including Airbus, where it handles billions of data points per day, B3 Exchange for its next-generation Central Securities Depository platform processing over $5 billion in daily traded volume, and BTG Pactual, where it serves as the standard for data aggregation with sub-10ms API response times and support for over 100,000 tickers per table.
QuestDB's key capabilities are backed by specific performance benchmarks. It ingests 8 million rows per second using the InfluxDB Line Protocol, and in published benchmarks it outperforms ClickHouse by 4.1-4.8x on ingestion, 27x on last-point queries, 3.4-4.9x on short-range point lookups, 1.3x on heavy scans, 2.1-2.3x on 12-hour range queries, and 1.2x on some double-groupby aggregations. It supports time-series-specific SQL constructs such as LATEST ON and SAMPLE BY, and its multi-tier storage engine auto-tiers data from hot memory to cold Parquet in object storage (S3, Azure, GCS, NFS, HDFS), with a query planner that spans partitions and Parquet files. The Enterprise edition adds a Rust-based infrastructure layer for high availability with replication and automatic failover, sub-second read replicas across regions, and elastic scaling with storage-compute separation. Security features include TLS, SSO via OAuth 2.0 and OIDC, role-based access control (RBAC), and audit logs, backed by a 99.9% uptime guarantee.
In the time-series database market, QuestDB competes directly with ClickHouse, TimescaleDB, and InfluxDB. Its published benchmarks show significant performance advantages over ClickHouse in ingestion and query speed, particularly for last-point and short-range queries, though ClickHouse may offer broader ecosystem integration and more mature distributed query capabilities. TimescaleDB and InfluxDB are more established in IoT and DevOps monitoring, but QuestDB claims 10x faster ingest than the next-fastest database in its class. The database is also compared to Kdb+, a legacy time-series database from KX, with QuestDB offering a SQL interface that lowers the barrier for adoption. QuestDB is deployed on trading floors and in mission control environments, and its open-source model with a commercial Enterprise tier positions it as a modern alternative to proprietary solutions.
Honest trade-offs exist. A notable limitation is that if any process opens a QuestDB table for writing, no other process—even in read-only mode—can open that table concurrently, which constrains multi-process architectures. The database is optimized for time-series workloads and may not be ideal for general-purpose OLTP or complex relational queries. The Enterprise edition, which provides high availability, elastic scaling, and SLA-backed support, requires a paid subscription starting at $65 per month, while the free version lacks replication and failover. The community edition is lightweight and designed to work out of the box, but users may need to invest in operational expertise for large-scale deployments, particularly when integrating with object storage and managing tiered storage policies. Finally, while QuestDB supports SQL and Parquet for data portability, its ecosystem of connectors and integrations is smaller than that of more mature databases like ClickHouse or InfluxDB.
How it works
-
Ultra-low latency queries
Delivers sub-millisecond query responses for time-series data, leveraging SIMD instructions and zero-GC Java/C++ to minimize latency.
-
High ingestion throughput
Ingests 8 million rows per second via the InfluxDB Line Protocol, with benchmarks showing 4.1-4.8x faster ingestion than ClickHouse.
-
Multi-tier storage engine
Auto-tiers data from hot memory to cold Parquet in object storage (S3, Azure, GCS), with a unified SQL planner spanning both tiers.
-
Native Parquet and SQL support
Stores cold data in open Parquet format and queries all tiers with standard SQL, avoiding vendor lock-in and enabling AI agent access.
-
Time-series SQL constructs
Provides LATEST ON and SAMPLE BY keywords for efficient last-point and aggregation queries, reducing query complexity for time-series patterns.
-
High availability and failover
Enterprise edition offers automatic replication, sub-second read replicas across regions, and 99.9% uptime guarantee with multi-AZ failover.
-
Security and governance
Includes TLS encryption, SSO via OAuth 2.0/OIDC, role-based access control, and audit logs for enterprise compliance.
Strengths and trade-offs
Strengths
- Benchmarks show QuestDB ingests data 4.1-4.8x faster than ClickHouse and 10x faster than the next-fastest time-series database.
- Last-point queries are 27x faster than ClickHouse, critical for real-time monitoring and trading applications.
- Supports over 100,000 tickers per table with sub-10ms API response times, as demonstrated at BTG Pactual.
- Deployed at tier-1 firms like Airbus, B3 Exchange, and BTG Pactual, processing billions of data points per day with 99.9% uptime.
Trade-offs
- If any process opens a table for writing, no other process—even in read-only mode—can access that table concurrently.
- The free version lacks high availability, replication, and automatic failover, which are only available in the paid Enterprise edition.
- Optimized primarily for time-series workloads, making it less suitable for general-purpose OLTP or complex relational queries.
- Ecosystem of connectors and integrations is smaller than more mature databases like ClickHouse or InfluxDB, potentially requiring custom development.
Pricing context
Free version available with community support. Paid plans start at $65 per month for Enterprise edition, which includes high availability, replication, SSO/RBAC, tiered storage, and SLA-backed support. Free trial available.
Getting started with QuestDB
-
Sign up for QuestDB
Go to questdb.io and click Get Started. Choose the free Community edition or start a 30-day trial of the Enterprise edition. Download the binary or use Docker to run QuestDB locally.
-
Connect your data source
Configure QuestDB to ingest streaming data using the InfluxDB Line Protocol. Set up your data pipeline to send metrics or events to the QuestDB HTTP endpoint or TCP port.
-
Define your table schema
Create a table with a designated timestamp column using SQL. Use the CREATE TABLE statement with the timestamp keyword to enable time-series optimizations like partitioning and SAMPLE BY.
-
Run a time-series query
Execute a SAMPLE BY query to aggregate data over time intervals, such as calculating average values per minute. Use LATEST ON to retrieve the most recent data point for each series.
-
Set up tiered storage
Configure the multi-tier storage engine by specifying a Parquet directory in object storage (S3, Azure, GCS). QuestDB automatically moves cold data from memory to Parquet, and you can query both tiers with SQL.
Frequently Asked Questions
What is QuestDB and what is it used for?
QuestDB is an open-source time-series database designed for ultra-low-latency and high-throughput workloads. It is used in latency-sensitive environments like trading floors, mission control centers, and real-time analytics in aerospace, defense, energy, and robotics.
How fast is QuestDB at ingesting data?
QuestDB ingests 8 million rows per second using the InfluxDB Line Protocol. Published benchmarks show it outperforms ClickHouse by 4.1-4.8x on ingestion and is 10x faster than the next-fastest time-series database in its class.
What is QuestDB's multi-tier storage engine?
QuestDB's multi-tier storage engine automatically tiers hot data in memory and cold data to Parquet files in object storage like S3, Azure, or GCS. It provides a unified SQL query surface across both tiers for seamless access.
How does QuestDB compare to ClickHouse?
QuestDB benchmarks show it is 4.1-4.8x faster on ingestion, 27x faster on last-point queries, and 3.4-4.9x faster on short-range point lookups than ClickHouse. However, ClickHouse may offer broader ecosystem integration and more mature distributed query capabilities.
What are the limitations of QuestDB?
A key limitation is that if any process opens a table for writing, no other process can access that table concurrently. The free version lacks high availability and replication. QuestDB is optimized for time-series workloads, not general-purpose OLTP.
How much does QuestDB Enterprise cost?
QuestDB Enterprise starts at $65 per month and includes high availability with replication, automatic failover, SSO/RBAC, tiered storage, and SLA-backed support. A free version with community support is also available.
Alternatives
How QuestDB compares
Direct head-to-head against 2 competitors. Picked by 7wData.
QuestDB
- Pricing
- Free version available with community support. Paid plans start at $65 per month for Enterprise edition, which includes high availability, replication, SSO/RBAC, tiered storage, and SLA-backed support. Free trial available.
- Target
- QuestDB is an open-source time-series database management system engineered for ultra-low-latency and high-throughput workloads.
- Strength
- Benchmarks show QuestDB ingests data 4.1-4.8x faster than ClickHouse and 10x faster than the next-fastest time-series database.
- Watch for
- If any process opens a table for writing, no other process—even in read-only mode—can access that table concurrently.
Google BigQuery
- Pricing
- $6.25/TiB after 1TiB free
- Target
- Enterprise analytics at scale
- Deployment
- Cloud-only
- Strength
- Petabyte SQL queries
- Watch for
- Cost spikes on complex joins
Redis
- Pricing
- Free tier + enterprise pricing
- Target
- Low-latency caching
- Deployment
- Any infra
- Strength
- Sub-millisecond reads
- Watch for
- Memory limits on time-series
User reviews
No user reviews yet. Be the first to write one.
Sources
Reporting on this tool draws on these publicly available sources.