dragonflydb

dragonfly: Redis and Memcached-Compatible In-Memory Store

Dragonfly is a high-throughput in-memory data store for modern application workloads, built on shared-nothing architecture and a Dash-based hash table. Its Redis and Memcached API compatibility is designed to let teams adopt it without changing application code.

Stars
31.8k
Forks
1.3k
Stars, last 30 days
+93
Last commit
today
Contributors
100+

Our take

Activeregular commits and releases

Health score

71/100

How it is scored
Activity
76
Community
87
Issues
64
Pull requests
58

+93 stars in the last 30 days

  • Merges more pull requests than 83% of the projects we track
  • More contributors than 78% of the projects we track

Dragonfly is an actively maintained, broadly contributed C++ datastore with a focused compatibility story, though teams should validate its command coverage and operating requirements against their workloads.

Good fit if

  • You need an in-memory store compatible with Redis or Memcached APIs and want to assess a replacement without application code changes.
  • Your workload can benefit from a shared-nothing design, built-in cache mode, or the included snapshot and metrics features.
  • You value signs of active maintenance: recent commits and releases, contributions from many people, and quick median issue and pull request handling.

Look elsewhere if

  • You require a specific Redis or Memcached command or behavior that Dragonfly does not support. The README describes roughly Redis 5.0-equivalent command coverage and a smaller Memcached command set, so verify your dependencies.
  • You need a license status established from the repository metadata. The supplied GitHub data reports NOASSERTION, so review the repository's license terms before adopting it.
  • Your application does not need an in-memory datastore or compatible cache API.
All health signals
Last commit2026-10-22 (today)
Commits, last 90 days100+
Releases, last 12 months12 (latest v2.0.2, 2026-10-07)
Contributors100+ (top contributor: 33% of commits)
Issues closed, last 90 days12+ (typically closed in 3 days)
Pull requests merged, last 90 days72 (typically merged in 1 day)
Project age4 years

Checked on 2026-10-23 with the GitHub API.

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

31,875

Forks

1,294

Change since Nov 14, 2026: +6 stars·0 forks

Overview

Dragonfly is an in-memory data store for modern application workloads, built on shared-nothing architecture and a hash table based on Dash. It aims to use CPU, memory, and I/O resources more effectively than legacy in-memory stores while preserving atomicity and low latency.

Its Redis and Memcached API compatibility makes it a candidate for teams looking to replace or scale an existing cache or key-value store without changing application code. The project also includes its own caching mode, snapshots, and a Prometheus-compatible metrics endpoint.

Key Features

  • Redis- and Memcached-compatible APIs, with no application code changes required for adoption.
  • Shared-nothing architecture that partitions the keyspace across threads.
  • Atomic multi-key operations supported by its transactional framework.
  • Cache mode with adaptive eviction when the configured memory limit is approached.
  • Snapshotting designed to avoid a fork-based process model.
  • Redis-style expiration and support for multiple databases.
  • Prometheus-compatible metrics and an HTTP console.
  • Cluster mode options and a Memcached-compatible API that can be enabled by configuration.

Use Cases

  • Teams running Redis-backed caches that need a store designed to scale across available CPU cores.
  • Applications using Memcached that want to evaluate a compatible alternative for read and write workloads.
  • Services that need an in-memory key-value store with atomic multi-key operations.
  • Operators who want cache eviction, snapshots, and metrics within the same datastore process.

What you need

Detected in the repository

  • Python >=3.10 (from pyproject.toml)
  • A test suite and automated checks on GitHub Actions

Getting Started

The README links to the Quick Start guide. Follow it for installation and launch instructions, then consult the repository README for configuration details.

License in Plain Words

No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.

Alternatives

  • KeyDB: KeyDB is a multithreaded Redis fork with features such as active replication, while Dragonfly uses a shared-nothing architecture and also supports Memcached APIs.
ProjectLanguageLicenseStarsStatus
dragonflyC++–31.8kActive
KeyDBC++BSD-3-Clause12.5kInactive

Considerations

Dragonfly's compatibility is broad but not unlimited: the README reports roughly 185 Redis commands, described as approximately equivalent to the Redis 5.0 API, and 13 Memcached commands. Check the supported command set and behavioral differences before switching production workloads. The README also notes that long millisecond-precision expiration deadlines are rounded to the nearest second, and that the HTTP console should only be used on a safe network or disabled when the main TCP port is externally exposed. The supplied license metadata is NOASSERTION, so confirm the applicable terms directly in the repository.

Found this useful?

Share it with someone who would like dragonfly.

Comparisons