QuestDB

What Is QuestDB?

Most time-series databases ask you to learn a proprietary query language, accept mediocre write speeds, or bolt time-series logic onto a general-purpose engine that was never designed for it. QuestDB rejects all three compromises. Built from scratch in zero-GC Java and C++, it is a column-oriented database purpose-built for time-series workloads — and it is fast. Benchmarks using the Time Series Benchmark Suite (TSBS) show QuestDB ingesting data 12-36x faster than InfluxDB and answering complex analytical queries 43-418x faster.

Founded in London in 2019 by a team with backgrounds in low-latency trading systems, QuestDB went through Y Combinator's Summer 2020 cohort and raised $14.3 million through a 2021 Series A led by 468 Capital, with angel investors including Reddit co-founder Alexis Ohanian. The company is small — roughly 20-30 employees — but has attracted a committed open-source community with over 14,000 GitHub stars.

The pitch is straightforward: standard SQL (not Flux, not PromQL) with purpose-built time-series extensions, running on an engine optimised at the instruction set level with SIMD vectorisation. You write SELECT * FROM trades SAMPLE BY 1h instead of learning a DSL. For teams already fluent in SQL, the migration cost is close to zero.

Key Features

SIMD-Accelerated Query Engine

QuestDB's core differentiator is raw speed. The engine uses Single Instruction Multiple Data (SIMD) instructions to process columnar data in parallel, combined with a custom JIT compiler that generates optimised machine code for filter expressions at query time. The result: sub-millisecond query latency on datasets that would take seconds in PostgreSQL-based alternatives. This is not theoretical — financial trading firms and IoT platforms depend on it in production.

SQL with Time-Series Extensions

Rather than inventing a new language, QuestDB extends standard SQL with operators designed for temporal data. SAMPLE BY aggregates across time windows. LATEST ON retrieves the most recent value per symbol — critical for tick data and sensor readings. ASOF JOIN merges datasets by closest timestamp, and the newer WINDOW JOIN and HORIZON JOIN handle more complex temporal relationships. If you know SQL, you already know 90% of QuestDB's query surface.

Multi-Tier Storage Engine

QuestDB 8.x introduced a three-tier storage architecture. Hot data enters via the Write-Ahead Log (WAL), moves to native columnar storage for fast querying, and automatically tiers to Apache Parquet on object storage (S3, Azure Blob, GCS) as it ages. Cold data in Parquet remains queryable in-place through QuestDB or any Parquet-compatible tool — Pandas, Polars, Spark. This eliminates vendor lock-in at the storage layer and keeps costs predictable as datasets grow into the terabytes.

PostgreSQL Wire Protocol

QuestDB speaks PostgreSQL's wire protocol on port 8812. This means any PostgreSQL-compatible driver, ORM, or BI tool connects without modification. Grafana, Superset, Power BI, DBeaver — they all work out of the box. For ingestion, QuestDB also supports InfluxDB Line Protocol (ILP) over TCP and HTTP, with official client libraries in seven languages including Python, Java, Go, Rust, C/C++, Node.js, and .NET.

Enterprise Security and HA

QuestDB Enterprise adds the operational controls missing from the open-source edition: Role-Based Access Control with column-level granularity, TLS encryption on all protocols, SSO via OAuth/OIDC/Azure Entra ID, and high-availability replication with point-in-time recovery. These are table-stakes features for regulated industries, and their absence in the open-source version is a legitimate limitation for production deployments handling sensitive data.

Pricing

QuestDB's pricing is a two-tier story with a gap in the middle.

The open-source edition is free under Apache 2.0. You get the full database engine — ingestion, storage, querying, time-series extensions — with no artificial limitations on data volume, retention, or throughput. Community support comes through a public Slack workspace and Discourse forum.

Enterprise pricing is not published. QuestDB discontinued its managed cloud service and pivoted to a Bring Your Own Cloud (BYOC) model, where their team manages QuestDB operations inside your AWS or Azure account. Self-managed Enterprise licences are also available. Both require contacting sales for a quote.

This pricing opacity is a real drawback. Competitors like TimescaleDB publish transparent cloud pricing tiers, and InfluxDB offers self-serve plans. QuestDB's approach works for enterprise procurement cycles but creates friction for mid-size teams evaluating options. If you are running QuestDB in production and want RBAC or TLS, you either accept the open-source limitations or enter a sales conversation with no public reference point for cost.

EU Compliance & Privacy

QuestDB's compliance story hinges on self-hosting. The open-source edition deploys anywhere — spin up a container in Frankfurt, Amsterdam, or any EU data centre and your time-series data never leaves the jurisdiction. No third-party processors, no transatlantic data transfers, no reliance on Standard Contractual Clauses.

The Enterprise BYOC model extends this: QuestDB manages the database but data stays in your cloud account, in your chosen region. For organisations subject to GDPR, FINRA, or sector-specific data localisation requirements, this is architecturally sound.

However, QuestDB Limited is a UK company. Post-Brexit, the UK operates under a separate data protection framework (UK GDPR) with an EU adequacy decision currently in place. This is lower-friction than US-based alternatives but not equivalent to an EU-domiciled provider. There are no published SOC 2, ISO 27001, or other third-party certifications. Enterprise security features — RBAC, TLS, SSO — are paywalled behind the Enterprise licence, meaning the open-source edition has no built-in authentication or encryption.

Who It's Best For

Quantitative finance and trading platforms needing nanosecond-precision timestamps, sub-millisecond query latency, and SQL-based analytics on tick data. QuestDB's heritage in low-latency systems shows here.

IoT and industrial telemetry teams ingesting millions of sensor readings per second. The combination of ILP ingestion and SIMD-accelerated queries handles high-cardinality time-series workloads that choke general-purpose databases.

DevOps and observability teams who want Grafana dashboards backed by a faster engine than Prometheus or InfluxDB, with the flexibility of SQL rather than PromQL.

Data engineering teams building modern data stacks where Parquet interoperability matters. QuestDB's cold-storage tier slots neatly into lakehouse architectures alongside tools like Spark, DuckDB, and Polars.

The Verdict

QuestDB earns its performance claims. For pure time-series ingestion and analytical query speed, nothing open-source touches it. The SQL interface makes it approachable, the Parquet tiering makes it economical, and self-hosting makes it deployable in any EU jurisdiction. The trade-offs are real: a small team, no published pricing for Enterprise, security features locked behind a commercial licence, and a UK domicile that adds a minor compliance footnote versus EU-headquartered alternatives. For teams where raw performance on time-series data is the primary requirement and self-hosting is acceptable, QuestDB is a compelling choice.

Frequently Asked Questions

Is QuestDB GDPR compliant?

The open-source edition is self-hosted, so compliance depends entirely on your deployment choices. Deploy in an EU data centre and data never leaves the jurisdiction. Enterprise BYOC keeps data in your own cloud account in your chosen region. QuestDB Limited is a UK company, operating under UK GDPR with an active EU adequacy decision.

How does QuestDB compare to InfluxDB?

QuestDB is significantly faster — 12-36x for ingestion, 43-418x for complex queries in TSBS benchmarks. QuestDB uses standard SQL while InfluxDB uses Flux (now deprecated) or SQL in v3. InfluxDB has a larger community and a self-serve cloud product. QuestDB has better raw performance but fewer managed deployment options.

Can I use QuestDB with Grafana?

Yes. QuestDB has an official Grafana plugin and also works through the standard PostgreSQL data source. Most teams use the official plugin for time-series specific features like SAMPLE BY queries in dashboard panels.

What happened to QuestDB Cloud?

QuestDB discontinued its managed cloud offering and transitioned to an Enterprise BYOC (Bring Your Own Cloud) model. Instead of running infrastructure on QuestDB's account, the database runs in your AWS or Azure account while QuestDB's team handles operations. This gives you more control over data residency but requires an existing cloud account.

Is QuestDB production-ready?

Yes. QuestDB is used in production by trading firms, IoT platforms, and observability teams processing billions of rows. The open-source edition is mature for workloads where built-in authentication is not required. For regulated environments needing RBAC, TLS, and replication, the Enterprise edition is necessary.