Why the Redis/Valkey debate is back in 2026
In March 2024, Redis Ltd. switched to the BSL (Business Source License), which is excluded from OSI open source definitions. The reaction was immediate: the Linux Foundation coordinated the creation of Valkey with support from Meta, Google, AWS and several major hosting providers. The fork started from the Redis 7.2 codebase under a BSD license. In late March 2026, Redis returned to an OSI-compatible license by adopting AGPLv3 — a positive signal, but AGPLv3 is a strong copyleft: any software that integrates it must publish its source code, which creates real constraints for reseller hosting providers and commercial software vendors. Meanwhile, Valkey 9.1 became the default package on Ubuntu 24.04+, Debian 13 and Fedora 40+, consolidating its adoption across major distributions.
Redis vs Valkey: key differences in 2026
| Criterion | Redis 8 (AGPLv3) | Valkey 9.1 (BSD) |
|---|---|---|
| License | AGPLv3 — strong copyleft, restricted commercial embedding | BSD — permissive, free commercial use |
| Performance | Historical reference | +8% ops/sec, -22% P99 latency, -20% memory |
| Protocol compatibility | 100% Redis protocol | 100% Redis protocol — transparent drop-in replacement |
| Distro support | Not installed by default | Ubuntu 24.04+, Debian 13, Fedora 40+ — native package |
| Modules | Redis Stack paid (RedisJSON, RediSearch…) | Compatible module API, growing ecosystem |
| Cloud adoption | AWS ElastiCache remains Redis | GCP Memorystore Valkey available, AWS moving toward Valkey |
| Best for | Existing projects depending on Redis Stack modules | New projects, hosting providers, sovereign stacks, modern distros |
When to choose Valkey over Redis
Valkey is the natural choice in several scenarios. Hosting providers and resellers are the most directly affected: Redis's AGPLv3 requires publishing the source code of any service that integrates it, a constraint incompatible with most commercial offerings. New projects have no reason to start with Redis: Valkey is the default package on recent distributions, performance is superior and the license is unrestricted. GDPR and sovereignty-focused stacks also benefit from the Linux Foundation's neutral governance compared to a private vendor. Finally, if you don't use Redis Stack proprietary modules (RedisJSON, RediSearch, RedisTimeSeries), migration is transparent — all your existing libraries and configurations remain valid.
Migrate from Redis to Valkey in 4 steps
Check compatibility
Run redis-cli INFO server | grep redis_version to identify your current version. List your dependencies and confirm none use Redis Stack modules (RedisJSON, RediSearch, RedisTimeSeries). If they do, direct migration is not yet recommended.
Install Valkey
On Ubuntu 24.04+ and Debian 13, Valkey is available in native repositories: apt install valkey. On Fedora 40+, use dnf install valkey. On older distributions, add the official Valkey repository before installing.
Copy your existing configuration
The configuration syntax is identical between Redis and Valkey. Simply copy your file: cp /etc/redis/redis.conf /etc/valkey/valkey.conf. Check socket and data paths if you have customized those settings.
Migrate data and validate
Export your Redis data via an RDB snapshot (redis-cli --rdb /tmp/dump.rdb) then import it into Valkey by placing the file in Valkey's dir directory. Start the service (systemctl start valkey), then validate with valkey-cli ping — you should get PONG. All your existing Redis libraries (ioredis, Predis, redis-py) are compatible without modification.
Valkey includes valkey-benchmark, functionally identical to redis-benchmark. Before any production migration, validate performance on your hardware with valkey-benchmark -n 100000 -q. This command sends 100,000 requests and displays an ops/sec summary for each command type. Expected result: figures higher than your previous Redis instance on the same server.
What about your other databases on VPS?
Valkey is just one piece of your data stack. If you also host a relational database, check our guide on MariaDB on VPS: configuration, performance and replication. For full-text search engines, Elasticsearch on VPS covers installation, memory tuning and production indexing. Each component deserves configuration tailored to your workload — our Cloud VPS gives you full control over your environment.
