PgCache vs Valkey

Valkey is an in-memory data store, forked from Redis in 2024, that many teams use to cache Postgres reads along with other jobs. PgCache is a caching proxy for Postgres. It overlaps with one of those jobs: caching query results.

Valkey facts checked against the Linux Foundation's announcement and Valkey's site on October 21, 2026.

What is Valkey?

Valkey is an in-memory data store and a fork of Redis. In March 2024, Redis Ltd. moved Redis from the BSD license to source-available licenses. About a week later, the Linux Foundation announced Valkey, built from Redis 7.2.4, the last BSD-licensed release. Valkey kept the BSD 3-Clause license. Amazon Web Services, Google Cloud, Oracle, Ericsson and Snap backed the project at launch.

Valkey started with the same data structures and commands as Redis, so most Redis clients work with it. The two projects have developed separately since the fork. Amazon ElastiCache offers Valkey as a managed service, alongside Redis OSS and Memcached.

What is PgCache?

PgCache is a caching proxy for PostgreSQL. Your app connects to PgCache instead of the database. PgCache stores the data for repeated read queries in a separate, vanilla Postgres instance and keeps that data current through Postgres logical replication. Writes, and reads PgCache can't serve from cache, go to your database.

PgCache works with PostgreSQL 16, 17 and 18, including Amazon RDS and Aurora. It's source-available under the Elastic License 2.0.

What Valkey is used for, and what PgCache covers

Use	PgCache	Valkey
Caching database query results	Yes, for Postgres, with no cache code in the app	Yes. The app stores results under keys and clears them when data changes.
Session storage	No	Yes
Pub/sub messaging	No	Yes
Queues and streams	No	Yes
Rate limits and counters	No	Yes
Leaderboards and rankings	No	Yes, with sorted sets

Where they differ

Topic	PgCache	Valkey
App changes	Point the connection string at PgCache	Cache code around each cached query and each write
What it caches	SQL query results, picked from live traffic	Whatever the app stores under a key
Freshness	Within typical logical replication lag	Until the app clears the key or its TTL expires
Writes from other services and tools	Reach the cache through replication	Not seen unless each one also clears the right keys
License	Elastic License 2.0 (source-available)	BSD 3-Clause

When each one fits

When PgCache fits

What you want to cache is Postgres query results
Several services or jobs write to the tables you cache

When Valkey fits

You need sessions, queues, pub/sub or other key-value data structures
You cache things that aren’t query results, like rendered pages or API responses

What they have in common

Both can serve repeated reads without querying Postgres
Neither speeds up writes

Using them together

Valkey for sessions, queues and other data that doesn’t live in Postgres
PgCache for caching Postgres query results

Common questions

Is Valkey the same as Redis?

Valkey started as a copy of Redis 7.2.4 in March 2024, and most Redis clients work with it. The two projects have developed separately since then.

Can PgCache replace Valkey?

It can replace the part of a Valkey setup that caches Postgres query results. Sessions, queues, pub/sub and Valkey’s other uses stay with Valkey.

Can I run Valkey and PgCache together?

Yes. PgCache sits between your app and Postgres and doesn’t touch Valkey.

Sources

If something here is out of date, please write to [email protected].

See how this applies to your database

The Fit Analyzer reads a pg_stat_statements export in your browser and shows which of your queries PgCache would serve. Questions are welcome at [email protected].