Last updated August 04, 2026

Stackhero for Valkey provides a managed Valkey instance running on a fully dedicated instance.

Valkey is an open-source fork of the Redis project before they changed their license. Valkey is a drop-in replacement for Redis 7.2 and previous versions.

With the Stackhero for Valkey add-on, you’ll get:

• A private instance (dedicated VM) for high performances and security
• TLS encryption (aka SSL)
• An automatic backup every 24 hours
• A graphical web UI (Redis Commander)
• One click to update to new Valkey versions
• Full-text search, JSON documents, vector search and Bloom filters included

All Redis clients can connect to Stackhero for Valkey, and there’s a Valkey/Redis client library for almost every language out there, including Ruby, Node.js, Java, Python, Clojure, and Erlang.

Provisioning the add-on

Stackhero for Valkey can be attached to a Heroku application via the CLI:

$ heroku addons:create --wait ah-valkey-stackhero --app <your app name>
Creating ah-valkey-stackhero...
Your add-on is being provisioned on Stackhero. It will be available in around 2 minutes.

After you provision Stackhero for Valkey, the STACKHERO_VALKEY_URL_TLS and STACKHERO_VALKEY_URL_CLEAR config variables are available in your app’s configuration. They contain the URLs to your Valkey instance as its credentials.

• STACKHERO_VALKEY_URL_TLS is the URL to your Valkey instance with TLS encryption. We recommend you use this URL to connect to your Valkey instance.
• STACKHERO_VALKEY_URL_CLEAR is the URL to your Valkey instance with no encryption (clear).

You can see the content of those variables via the heroku config:get command:

$ heroku config:get STACKHERO_VALKEY_URL_TLS
rediss://user:password@domain:port

After you install Stackhero for Valkey, you must configure your application to fully integrate with the add-on.

Local setup

After you provision the add-on, it’s necessary to locally replicate its config variables so your development environment can operate against the service.

Use the Heroku Local command-line tool to configure, run, and manage process types specified in your app’s Procfile. Heroku Local reads configuration variables from a .env file. To view all of your app’s config variables, type heroku config. Use these commands for each value that you want to add to your .env file:

$ heroku config:get STACKHERO_VALKEY_URL_TLS -s  >> .env
$ heroku config:get STACKHERO_VALKEY_URL_CLEAR -s  >> .env

For more information, see Running Apps Locally.

Configure Valkey on Java

You can use the environment variable STACKHERO_VALKEY_URL_TLS to connect to Valkey.

Here’s an example of a connection using Jedis:

private static Jedis getConnection() throws URISyntaxException {
  URI valkeyURI = new URI(System.getenv("STACKHERO_VALKEY_URL_TLS"));
  Jedis jedis = new Jedis(valkeyURI);
  return jedis;
}

For more information, see the official Jedis repository and the official wiki.

In a multithreaded environment, like a web server, use Jedis Pool:

public static JedisPool getPool() {
  URI valkeyURI = new URI(System.getenv("STACKHERO_VALKEY_URL_TLS"));
  JedisPoolConfig poolConfig = new JedisPoolConfig();
  poolConfig.setMaxTotal(10);
  poolConfig.setMaxIdle(5);
  poolConfig.setMinIdle(1);
  poolConfig.setTestOnBorrow(true);
  poolConfig.setTestOnReturn(true);
  poolConfig.setTestWhileIdle(true);
  JedisPool pool = new JedisPool(poolConfig, valkeyURI);
  return pool;
}

For more information, see the official Jedis wiki.

Configure Valkey on Ruby

Install the Redis gem:

$ bundle add redis

With Rails, you have to create the initializer file config/initializers/redis.rb like this:

$redis = Redis.new(url: ENV["STACKHERO_VALKEY_URL_TLS"])

To use Valkey as a cache system, edit the config/environments/production.rb file and add this line:

config.cache_store = :redis_cache_store, { url: ENV['STACKHERO_VALKEY_URL_TLS'] }

Configure Valkey on Sidekiq

To use Stackhero for Valkey as your Sidekiq Valkey server, set the environment variable REDIS_PROVIDER to STACKHERO_VALKEY_URL_TLS:

$ heroku config:set REDIS_PROVIDER=STACKHERO_VALKEY_URL_TLS

Sidekiq will automatically use Stackero for Valkey then.

Configure Valkey on Resque

Edit the file config/resque.yml and replace production: <%= ENV['REDIS_URL'] %> with this value:

production: <%= ENV['STACKHERO_VALKEY_URL_TLS'] %>

Resque will then use Stackero for Valkey.

Configure Valkey on Python

Install the Redis package:

$ pip install redis
$ pip freeze > requirements.txt

import os
import redis

r = redis.from_url(os.environ.get("STACKHERO_VALKEY_URL_TLS"))

For more information, see the official Python redis package documentation.

How to avoid error “Connection closed by server” with Valkey and Python

The error redis.exceptions.ConnectionError: Connection closed by server is potentially related to the fact that your Python app doesn’t exchange data with Valkey for a certain time and the connection closes automatically. When your app tries to exchange again with Valkey, the connection doesn’t work anymore and you get the error Connection closed by server.

To resolve this, you can pass the health_check_interval setting to your Valkey connection like this:

r = redis.from_url(
  'rediss://default:<password>@XXXXXX.stackhero-network.com:<port>',
  health_check_interval=10,
  socket_connect_timeout=5,
  retry_on_timeout=True,
  socket_keepalive=True
)

If you use the Valkey PubSub feature, the redis-py library assumes that you call functions get_message() or listen() more frequently than the health_check_interval value in seconds. In this example, health_check_interval is set to 10 seconds. This means your app has to call get_message() or listen() at least one time each 10 seconds. For more information, see the redis-py official documentation.

If this isn’t the case, you’ll get the error Connection closed by server.

To avoid that, a trick is to call check_health() regularly.

Here’s an example:

import redis
import threading

# Connection to Valkey
r = redis.from_url(
  'rediss://default:<password>@XXXXXX.stackhero-network.com:<port>',
  health_check_interval=10,
  socket_connect_timeout=5,
  retry_on_timeout=True,
  socket_keepalive=True
)

# Create a PubSub instance
p = r.pubsub()

# Subscribe to the channel "test"
p.subscribe('test')

# Create a function that will call `check_health` every 5 seconds
def redis_auto_check(p):
  t = threading.Timer(5, redis_auto_check, [ p ])
  t.start()
  p.check_health()

# Call the redis_auto_check function
redis_auto_check(p)

Configure Valkey on PHP

Retrieve the URL with getenv('STACKHERO_VALKEY_URL_TLS') and pass it to your preferred Valkey client library.

Handle PHP sessions with Valkey

You can use the following code to store PHP sessions on Stackhero for Valkey:

<?php

// Parse Heroku Valkey URL from environment
$valkey_url = parse_url(getenv('STACKHERO_VALKEY_URL_TLS'));

// Configure session handler
ini_set('session.save_handler', 'redis');
ini_set('session.save_path', "tls://{$valkey_url['host']}:{$valkey_url['port']}?auth={$valkey_url['pass']}&timeout=5");

// Start the session
session_start();

?>

Configure Valkey on Go

First install the go-redis package:

$ go get github.com/redis/go-redis/v9

Then, import it in your code:

import "github.com/redis/go-redis/v9"

Finally, connect to the Valkey server using the STACKHERO_VALKEY_URL_TLS variable:

opt, err := redis.ParseURL(os.Getenv("STACKHERO_VALKEY_URL_TLS"))
if err != nil {
    panic(err)
}

rdb := redis.NewClient(opt)

For more information, see the official go-redis repository

Configure Valkey on Node.js

We recommend using ioredis.

Install the ioredis package:

$ npm install ioredis

const Ioredis = require('ioredis');

(async () => {

  const valkey = new Ioredis(process.env.STACKHERO_VALKEY_URL_TLS);

  // Set key "stackhero-example-key" to "abcd"
  await valkey.set('stackhero-example-key', 'abcd');

  // Get key "stackhero-example-key"
  const value = await valkey.get('stackhero-example-key');
  console.log(`Key "stackhero-example-key" has value "${value}"`);

  // Finally delete key "stackhero-example-key"
  await valkey.del('stackhero-example-key');

})().catch(error => {
  console.error('An error occurred!', error);
});

For more examples, see the official ioredis repository.

Dashboard and web UI (Redis Commander)

Stackhero dashboard allows you to see your instance usage, restart it, and apply updates. It also gives you the ability to access the web UI to consult your Valkey data directly in a graphical way.

You can access the dashboard via the CLI:

$ heroku addons:open ah-valkey-stackhero
Opening ah-valkey-stackhero for sharp-mountain-4005

You can also visit the Heroku Dashboard and select the application in question. Then, select Stackhero for Valkey from the Add-ons menu.

Set eviction policy

Yon can define how Valkey will react when you consume more memory than available on your plan.

To do so, connect to your Stackhero dashboard, select your Valkey service, then click Configure and set the Eviction policy setting.

Full-text search, JSON documents and vector search

Your Stackhero for Valkey includes the modules maintained by the Valkey project, at no extra cost. They turn Valkey into much more than a cache: you can index and query your data without adding a search engine or a vector database next to it.

• JSON (JSON.* commands): store real JSON documents and read or update a single field with JSONPath, instead of serializing a whole object into a string.
• Search (FT.* commands): secondary indexes, full-text search, numeric and tag filters, and vector search.
• Bloom (BF.* commands): probabilistic structures that answer “have I already seen this?” over huge sets, in a few kilobytes.

Enabling the modules

Modules are disabled by default, so an existing instance keeps exactly the behaviour it had.

To enable them, connect to your Stackhero dashboard (see above), select your Valkey service, click on Configure, then tick the modules you need in the Modules section and save. Your Valkey restarts with them loaded, which takes a few seconds.

Storing and querying JSON documents

Valkey speaks the same protocol as Redis, so any Redis client works.

const Ioredis = require('ioredis');

(async () => {

  const valkey = new Ioredis(process.env.STACKHERO_VALKEY_URL_TLS);

  // Store a product as a real JSON document
  await valkey.call(
    'JSON.SET', 'product:1', '$',
    JSON.stringify({ name: 'Red running shoes', brand: 'Acme', price: 89.9, stock: 12 })
  );

  // Read a single field, without transferring the whole document
  const price = await valkey.call('JSON.GET', 'product:1', '$.price');
  console.log(`Price: ${price}`); // Price: [89.9]

  await valkey.quit();

})().catch(error => {
  console.error('An error occurred!', error);
});

Full-text search

An index is created once and then stays up to date on its own: every document matching its prefix is indexed as you write it.

// Create the index once
await valkey.call(
  'FT.CREATE', 'productsIndex',
  'ON', 'JSON',
  'PREFIX', '1', 'product:',
  'SCHEMA',
    '$.name', 'AS', 'name', 'TEXT',
    '$.brand', 'AS', 'brand', 'TAG',
    '$.price', 'AS', 'price', 'NUMERIC'
);

// Full-text search
console.log(
  await valkey.call('FT.SEARCH', 'productsIndex', 'running', 'RETURN', '1', 'name')
);

// Full-text, tag and numeric filters combined in a single query
console.log(
  await valkey.call(
    'FT.SEARCH', 'productsIndex', '@brand:{Acme} @price:[0 100]',
    'RETURN', '2', 'name', 'price'
  )
);

Vector search, for semantic search and RAG

If you are building semantic search, a recommendation engine or a retrieval-augmented generation (RAG) pipeline, you need to store embeddings and find the closest ones to a query. The search module does that natively, and it can combine a vector similarity search with regular filters in a single query: “the 5 chunks closest to this question, but only from the documents this user may read” is one request, not three.

// A vector field is declared with its dimension and its distance metric.
// The dimension must match your embeddings model: 1536 for OpenAI
// text-embedding-3-small, 768 for many open source models, and so on.
await valkey.call(
  'FT.CREATE', 'chunksIndex',
  'ON', 'HASH',
  'PREFIX', '1', 'chunk:',
  'SCHEMA',
    'content', 'TEXT',
    'documentId', 'TAG',
    'embedding', 'VECTOR', 'HNSW', '6',
      'TYPE', 'FLOAT32',
      'DIM', '1536',
      'DISTANCE_METRIC', 'COSINE'
);

// Embeddings are stored as raw little-endian float32 bytes
const toBytes = embedding => Buffer.from(Float32Array.from(embedding).buffer);

await valkey.call(
  'HSET', 'chunk:1',
  'content', 'Refunds are issued within 14 days',
  'documentId', 'faq',
  'embedding', toBytes(chunkEmbedding)
);

// The 5 chunks closest to the question, restricted to one document.
// Results come back ordered by distance, closest first.
const results = await valkey.call(
  'FT.SEARCH', 'chunksIndex',
  '@documentId:{faq}=>[KNN 5 @embedding $queryVector AS score]',
  'RETURN', '2', 'content', 'score',
  'PARAMS', '2', 'queryVector', toBytes(questionEmbedding)
);

Indexes consume the memory of your plan, exactly like your data does. A large text or vector index can be significant, so watch your memory usage on your Stackhero dashboard after building one, and move to a larger plan if you need to.

You will find complete guides on search and JSON and on vector search and RAG in the Valkey documentation by Stackhero.

Upgrading your plan

 

Use the heroku addons:upgrade command to migrate to a new plan.

$ heroku addons:upgrade ah-valkey-stackhero:newplan
-----> Upgrading ah-valkey-stackhero:newplan to sharp-mountain-4005... done
       Your plan has been updated to: ah-valkey-stackhero:newplan

Removing the add-on

You can remove Stackhero for Valkey via the CLI:

$ heroku addons:destroy ah-valkey-stackhero
-----> Removing ah-valkey-stackhero from sharp-mountain-4005... done

Support

You can submit support and runtime issues for Stackhero for Valkey via one of the Heroku Support channels. For urgent issues, CC support@stackhero.io.

Additional resources

• Valkey documentation by Stackhero
• Valkey managed cloud