Valkey

Valkey: Drop-In Redis Replacement with Better Performance and No Licensing Worries

Michael Soto

23 Dec 2025 • 4 min read

Look, if you're still running Redis in production and haven't heard about Valkey yet, we need to talk. Redis changed their license in 2024, and while that sent shockwaves through the community, something better emerged from the chaos: Valkey—a fully open-source fork that's not just a Redis clone, but actually faster.

Here's the thing: Valkey is 100% Redis-compatible. Your existing code, client libraries, and configurations work without modification. But under the hood, Valkey 8.x brings architectural improvements that Redis simply doesn't have. We're talking 37% higher throughput on writes and 30% faster p99 latencies.

Let's get you migrated.

Why Valkey Over Redis?

Three reasons:

1. It's actually open source. BSD-3-Clause license, community-driven, no licensing surprises. Major cloud providers like AWS already switched ElastiCache to Valkey by default.

2. Performance improvements matter. Valkey's async I/O threading model (contributed by AWS) pushes throughput from ~240K RPS to ~680K RPS with 6 I/O threads. That's nearly 3x the throughput while p99 latencies drop from 1.68ms to 0.93ms.

3. Active development. Valkey 8.0 already ships with multi-threading support, dual-channel replication, and better memory efficiency. Valkey 9.0 (coming 2025) brings atomic slot migration and clustering improvements that Redis hasn't implemented.

The Migration: Easier Than You Think

If you can copy a file, you can migrate to Valkey. Here's the straightforward approach:

Step 1: Create a Redis Snapshot

redis-cli BGSAVE

Wait for the background save to complete:

redis-cli LASTSAVE

Your snapshot lives at /var/lib/redis/dump.rdb (or wherever your Redis data directory is configured).

Step 2: Deploy Valkey

If you're self-hosting, here's a production-ready Docker Compose configuration:

version: '3.8'
services:
  valkey:
    image: valkey/valkey:8-alpine
    container_name: valkey
    ports:
      - "172.17.0.1:6379:6379"
    volumes:
      - ./data:/data
      - ./valkey.conf:/usr/local/etc/valkey/valkey.conf
    command: valkey-server /usr/local/etc/valkey/valkey.conf
    restart: always

Step 3: Copy Your Data

Stop your Redis instance, copy the dump.rdb file to Valkey's data directory:

# Stop Redis
redis-cli SHUTDOWN SAVE

# Copy snapshot to Valkey
cp /var/lib/redis/dump.rdb ./data/dump.rdb

# Start Valkey
docker-compose up -d

Valkey automatically loads the RDB file on startup. Your data is now in Valkey.

Step 4: Update Connection Strings

Here's the beautiful part—you probably don't need to change anything. Valkey speaks the same protocol. Just point your applications to the new Valkey endpoint:

# Before
REDIS_URL=redis://old-redis-server:6379

# After
REDIS_URL=redis://valkey-server:6379

Same clients, same commands, better performance.

Enabling Multi-Threading (The Performance Win)

Valkey's I/O threading is where the magic happens. Create a valkey.conf file:

# Enable I/O threading
io-threads 4
io-threads-do-reads yes

# Memory management
maxmemory 2gb
maxmemory-policy allkeys-lru

# Persistence
appendonly yes
appendfsync everysec

# Security (set your password)
requirepass your-secure-password

The rule of thumb: set io-threads to roughly half your CPU cores. On an 8-core machine, 4 I/O threads typically deliver optimal throughput without context-switching overhead.

Quick Benchmark

After migration, verify you're getting the performance gains. Run valkey-benchmark:

docker exec -it valkey valkey-benchmark -h 127.0.0.1 -p 6379 \
  -t set,get -n 100000 -c 50 --threads 4

You should see throughput numbers significantly higher than your Redis baseline, especially with I/O threading enabled.

Configuration Gotchas

A few things to watch:

1. ACL compatibility: Valkey maintains Redis 7.2.4 ACL compatibility. If you're running Redis 7.x ACLs, they'll work unchanged.

2. Module compatibility: Most Redis modules work, but check compatibility for any custom modules you're using. Core data structures (strings, hashes, lists, sets, sorted sets) are 100% compatible.

3. Replication: If you're running Redis Cluster, Valkey supports the same cluster protocol. You can even run Valkey replicas against Redis primaries during transition.

Troubleshooting

Issue: Valkey won't start with my RDB file

Check your Redis version. Valkey forked from Redis 7.2.4, so if you're on Redis 6.x, upgrade to 7.2 first, create a new snapshot, then migrate.

redis-cli INFO server | grep redis_version

Issue: Performance isn't better

Enable I/O threading. Without it, you're essentially running single-threaded like Redis. Also check that your client supports pipelining—Valkey shines with concurrent connections.

Issue: Memory usage seems higher

Valkey 8.x actually improved memory efficiency with per-slot dictionaries in cluster mode. If you're seeing higher usage, check your maxmemory policy and ensure you're not running with persistence disabled (which can balloon memory during heavy writes).

Deploy on Elestio

Don't want to manage infrastructure? Elestio offers managed Valkey with:

Automated backups and updates
Built-in monitoring
Multi-cloud deployment (Hetzner, DigitalOcean, AWS, Vultr)
Starting at ~$16/month for a 2 CPU / 4GB instance

You get the performance benefits of Valkey without the operational overhead. They also offer database migration services if you need help moving from an existing Redis deployment.

The Bottom Line

Migrating from Redis to Valkey is a 30-minute operation that delivers measurable performance improvements and eliminates licensing concerns. The community has spoken—Valkey is the future of open-source in-memory data stores.

Your code doesn't change. Your clients don't change. Your data structure doesn't change. What changes is throughput, latency, and peace of mind.

If you're still on Redis, there's genuinely no reason to wait.

Thanks for reading!