Dragonfly
dragonflydb.io
- Free trial
- Linux
- API
- Self-hosted
What Dragonfly is
Dragonfly is an in-memory data store for real-time workloads such as caching, feature serving, message queues, and sorted-set leaderboards. It supports strings, hashes, lists, sets, sorted sets, streams, JSON, vectors, geospatial data, and probabilistic structures. Data persistence, high availability, and vector search are supported; SQL support is not. Redis and Memcached APIs are compatible, and the maker says adoption needs no code changes. Its multi-threaded, shared-nothing architecture spreads work across CPU cores and automatically shards data. Dragonfly states it can deliver millions of queries per second per instance with sub-1ms P99 latency, while data structures and compression improve memory utilization by 30–60%. It supports agentless Prometheus monitoring and S3-compatible snapshots. Docker, Kubernetes, cloud, and on-premises deployments are available. Dragonfly Cloud starts at $36 a month.
Who it suits
It suits teams needing an in-memory store for caching, queues, feature serving, or leaderboard workloads. Redis or Memcached API compatibility may help existing deployments adopt it.
Good to know, and what to watch
- Redis and Memcached API compatibility
- Automatic sharding across CPU cores
- Supports persistence and high availability
- Agentless Prometheus monitoring
- Runs in cloud or on-premises
- No free plan is available
- Business starts at $2,000 monthly
Plans
Every plan and priceSelf-Managed
Flex
Business
The verdict
Dragonfly combines familiar APIs with multi-core processing, broad data types, and several deployment options. It is a paid choice for real-time workloads where persistence and high availability are required.
What the maker publishes
- API compatibility
- Dragonfly is compatible with Redis and Memcached APIs, and the maker says adoption requires no code changes. source
- Architecture
- Dragonfly uses a multi-threaded, shared-nothing architecture that processes work across CPU cores. source
- Cloud hosting
- Dragonfly Cloud is managed on AWS, GCP, Azure, or on-premises, with a bring-your-own-cloud option that keeps data in the customer’s account. source
- Company headquarters
- LinkedIn lists Dragonfly’s headquarters as Tel Aviv. source
- Company history
- DragonflyDB says the company was founded in 2022. source
- Deployment
- Dragonfly can be run locally via Docker, and its single binary can be deployed via Docker, Kubernetes, or elsewhere. source
- Integrations
- Dragonfly supports Prometheus monitoring without an agent and can read and write snapshots to S3-compatible storage. source
- Memory efficiency
- The maker says Dragonfly offers 30–60% better memory utilization through its data structures and compression. source
- Operations
- Dragonfly automatically shards data between CPUs, eliminating the need to reshard. source
- Performance
- The features page states Dragonfly can deliver millions of queries per second per instance and sub-1ms P99 latency. source
- Product
- Dragonfly is a modern in-memory data store for real-time workloads, caching, feature stores, and job queues. source
- Security
- Dragonfly Cloud lists RBAC, ACL, user access management, and administrative audit logs as security controls. source
- Support
- Dragonfly Cloud offers support from its engineering team backed by contractual SLAs; its pricing page also lists dedicated support for Business. source
- Use cases
- The maker highlights caching, real-time feature serving, message queues, and sorted-set leaderboards as use cases. source
Checked against the maker's pages on 3 Oct 2026.