Getting started

1. Create an instance

Sign in with Google or GitHub, open the dashboard and click New instance. Pick a name, an engine and a plan. The instance is usually ready in under a minute; the dashboard updates by itself.

2. Connection details

The instance page in the dashboard shows everything you need, with copy buttons:

Note the double s in rediss://: it tells clients to use TLS. The examples in these docs read the URL from an environment variable:

export VALKEY_URL='rediss://default:<password>@inst-<id>.eu.kvnest.com:6380'

3. Try it from your terminal

With redis-cli 6 or newer, or valkey-cli, which should answer PONG. Either works with either engine:

redis-cli --tls --sni inst-<id>.eu.kvnest.com \  -u 'rediss://default:<password>@inst-<id>.eu.kvnest.com:6380' PING

valkey-cli --tls --sni inst-<id>.eu.kvnest.com \  -h inst-<id>.eu.kvnest.com -p 6380 \  --user default --pass '<password>' PING

4. Connect from your app

Pick your language for client libraries and framework integrations:

5. Automate it

Manage instances without the dashboard, with a personal API token:

Limits and behaviour

• Memory and connections follow your plan. Starter has 192 MB usable memory and up to 100 clients; keep connection pool sizes below that across all your app instances.
• Starter is a cache. It writes nothing to disk and, when full, evicts the least recently used keys (allkeys-lru). Data is lost when the instance restarts. Don't use it as your only copy of anything, or for job queues.
• Eviction policy defaults to the plan's and can be changed on the instance's Settings tab, for example to noeviction so writes fail instead of dropping keys when memory is full. Changing it restarts the instance.
• Modules: turn on JSON (JSON.*), Bloom filters (BF.*) and Search (FT.*, including vector similarity search) when creating an instance or on its Settings tab; Redis instances also get TimeSeries (TS.*). Search indexes count towards the memory limit. Switching modules restarts the instance.
• No backups during the preview, on any plan.
• Idle connections are closed after 5 minutes. Pooled clients reconnect on their own; long-lived single connections should enable reconnects or a periodic health check.
• Disabled commands:CONFIG, DEBUG, MODULE, SHUTDOWN, REPLICAOF, SLAVEOF, FAILOVER, SYNC, PSYNC. Libraries that call CONFIG have an option to skip it, for example Spring Session's spring.session.redis.configure-action=none.
• Engines: choose Redis Open Source 8.10 or Valkey 9.1 when you create an instance; it can't be switched later. Both speak the same protocol, so any Redis or Valkey client works with either. Which one should I pick?

Redis or Valkey?

Both are first-class on KVNest: same plans, prices, TLS, users and metrics. Valkey is the Linux Foundation's fork of Redis 7.2, under the BSD licence. Redis Open Source 8 adds new commands and data types on top, and is offered under the AGPLv3; KVNest runs it unmodified.

Not sure? Either works for caching, sessions, queues and leaderboards, and your code doesn't change between them.

Troubleshooting