DEV.co
Open-Source Databases · polarsignals

frostdb

FrostDB is an embeddable columnar database written in Go, designed for write-heavy observability workloads. It uses Apache Parquet for storage and Apache Arrow for queries, with support for dynamic schemas that adapt at runtime.

Source: GitHub — github.com/polarsignals/frostdb
1.5k
GitHub stars
76
Forks
Go
Primary language
Apache-2.0
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositorypolarsignals/frostdb
Ownerpolarsignals
Primary languageGo
LicenseApache-2.0 — OSI-approved
Stars1.5k
Forks76
Open issues18
Latest releaseUnknown
Last updated2026-07-08
Sourcehttps://github.com/polarsignals/frostdb

What frostdb is

FrostDB implements a columnar storage engine with LSM-tree compaction, snapshot isolation via transaction watermarking, and dynamic column support for wide-column schemas. It optimizes for immutable data, leveraging Parquet sparse indices (bloom filters, min/max stats) to reduce query overhead.

Quickstart

Get the frostdb source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/polarsignals/frostdb.gitcd frostdb# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Observability & Metrics Storage

Designed for Prometheus-like time-series with dynamic labels. High-cardinality dimensions and frequent aggregations benefit from columnar layout; LSM compaction suits append-heavy observability pipelines.

Embedded Analytics in Go Services

In-process database avoids separate server deployment. Suitable for applications needing local analytical queries over write-heavy datasets without distributed complexity.

Immutable, Append-Only Data Workloads

Optimized for write-once, many-read patterns. Event logging, audit trails, and time-series data that never require updates or deletes perform well with snapshot isolation semantics.

Implementation considerations

  • No latest release tag in repository; assess HEAD stability and test thoroughly before committing to a dependency version.
  • Dynamic schema requires careful field naming conventions (e.g., `frostdb:` tags); schema design choices will impact query performance and storage efficiency.
  • Snapshot isolation lacks read-after-write consistency—readers must re-query to see new data; application code must handle eventual consistency.
  • LSM compaction and index rotation are configurable; tuning in-memory levels, size thresholds, and storage backends is critical for your workload.
  • Immutability simplifies concurrency but constrains use cases; confirm append-only semantics align with your data model before deep integration.

When to avoid it — and what to weigh

  • Requires Update/Delete Operations — FrostDB is immutable-only. Row modifications, partial deletes, or in-place updates are not supported. Workloads needing transactional CRUD are unsuitable.
  • Non-Go Ecosystem or Standalone Server — FrostDB is embedded-only and Go-native. Projects in Python, Java, or Node.js, or those requiring a shared database server with network access, should look elsewhere.
  • Row-Oriented Query Patterns — FrostDB excels at column scans and aggregations. If your workload is primarily row-level point lookups (get row by ID), columnar overhead will hurt performance.
  • Production-Ready Guarantee Required — README explicitly states: 'still in its infancy... not production-ready, probably has various consistency and correctness problems and all API will change.' Risk of breaking changes.

License & commercial use

Licensed under Apache License 2.0 (Apache-2.0), a permissive OSI-approved license. Allows commercial use, modification, and distribution with attribution and liability disclaimers.

Apache-2.0 is a permissive license that allows commercial use. However, given the project's stated pre-release status (not production-ready, API will change), commercial deployment carries significant risk. Ensure your legal and engineering teams review the stability guarantees and support expectations before relying on this in production systems.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationAdequate
License clarityClear
Deployment complexityModerate
DEV.co fitGood
Assessment confidenceMedium
Security considerations

No explicit security audit, vulnerability disclosure process, or threat model documented. Immutability and local-only (embedded) deployment reduce some attack surfaces, but encryption at rest, encryption in transit, and access control are not mentioned. Security posture requires independent review before handling sensitive data.

Alternatives to consider

DuckDB

Similar embeddable columnar database; supports SQL and multiple languages (Python, R, C++); more mature and battle-tested; better for general analytics. Downside: not Go-native, less optimized for dynamic schemas.

ClickHouse

High-performance columnar OLAP engine; excellent for observability (Prometheus integration exists); standalone server model. Downside: requires separate deployment, heavier operational footprint, not embeddable.

Prometheus + Local Storage

Purpose-built for metrics; simpler schema and query model; battle-hardened in production observability. Downside: limited to time-series; storage layer not designed for broad analytics or dynamic columns beyond labels.

Software development agency

Build on frostdb with DEV.co software developers

FrostDB is powerful for observability and immutable workloads but pre-release. Our engineers can help you assess fit, prototype integration, and plan migration. Contact us for a technical review.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

frostdb FAQ

Is FrostDB production-ready?
No. The README explicitly warns: 'not production-ready, probably has various consistency and correctness problems and all API will change.' Use only for prototyping, proof-of-concept, or non-critical systems.
Can I update or delete data once written?
No. FrostDB is immutable-only. All data is append-only; updates and deletes are not supported. Workloads requiring modifications need a different database.
Does FrostDB support SQL?
Not explicitly mentioned in the README. It provides a DataFrame-like query builder API; SQL support status is unknown and should be verified in the codebase or documentation.
How do I persist data to cloud storage?
Persistence to Parquet files is configurable, but integration with S3, GCS, or other cloud services is not documented. Custom implementation required.

Work with a software development agency

DEV.co is a software development agency delivering custom software development services to companies building on open source. Our software developers and web developers design, integrate, and ship production systems — spanning web development, APIs, AI, data, and cloud. If frostdb is part of your open-source databases roadmap, our team can implement, customize, migrate, and maintain it.

Evaluating FrostDB for Your Stack?

FrostDB is powerful for observability and immutable workloads but pre-release. Our engineers can help you assess fit, prototype integration, and plan migration. Contact us for a technical review.