Redis store for kemal-session
HEAD Latest release released



Redis session store for kemal-session.


Add this to your application's shard.yml:

    github: neovintage/kemal-session-redis
    version: 1.0.1


require "kemal"
require "kemal-session"
require "kemal-session-redis"

Kemal::Session.config do |config|
  config.cookie_name = "redis_test"
  config.secret = "a_secret"
  config.engine = "localhost", port: 1234)
  config.timeout =, 0, 0)

get "/" do
  puts "Hello World"

post "/sign_in" do |context|"see-it-works", 1)

The engine comes with a number of configuration options:

| Option | Description | | ------ | ----------- | | host | where your redis instance lives | | port | assigned port for redis instance | | unixsocket | Use a socket instead of host/port. This will override host / port settings | | database | which database to use when after connecting to redis. defaults to 0 | | capacity | how many connections the connection pool should create. defaults to 20 | | timeout | how long until a connection is considered long-running. defaults to 2.0 (seconds) | | key_prefix | when saving sessions to redis, how should the keys be namespaced. defaults to kemal:session: |

When the Redis engine is instantiated, RedisEngine will create a connection pool for you. The pool will have 20 connections and a timeout of 2 seconds.

If no options are passed the RedisEngine will try to connect to a Redis using default settings.

Best Practices

Creating a Client

It's very easy for client code to leak Redis connections and you should pass a pool of connections that's used throughout Kemal and the session engine.

Session Administration Performance

Kemal::Session.all and Kemal::Session.each perform a bit differently under the hood. If Kemal::Session.all is used, the RedisEngine will use the SCAN command in Redis and page through all of the sessions, hydrating the Session object and returing an array of all sessions. If session storage has a large number of sessions this could have performance implications. Kemal::Session.each also uses the SCAN command in Redis but instead of creating one large array and enumerating through it, Kemal::Session.each will only hydrate and yield the keys returned from the current cursor. Once that block of sessions has been yielded, RedisEngine will retrieve the next block of sessions.


Redis must be running on localhost and bound to the default port to run specs.


  1. Fork it ( )
  2. Create your feature branch (git checkout -b my-new-feature)
  3. Commit your changes (git commit -am 'Add some feature')
  4. Push to the branch (git push origin my-new-feature)
  5. Create a new Pull Request


  github: mamantoha/kemal-session-redis
License MIT
Crystal none


Dependencies 2

  • kemal-session ~> 1.0
    {'github' => 'kemalcr/kemal-session', 'version' => '~> 1.0'}
  • redis ~> 2.8.3
    {'github' => 'stefanwille/crystal-redis', 'version' => '~> 2.8.3'}

Development Dependencies 0

Dependents 0

Last synced .
search fire star recently