The elastic-transport library provides a low-level Ruby client for connecting to an Elasticsearch cluster. It currently powers the Elasticsearch Ruby and the Enterprise Search Ruby clients.

When available, it handles connecting to multiple nodes in the cluster, rotating across connections, logging and tracing requests and responses, maintaining failed connections, discovering nodes in the cluster, and provides an abstraction for data serialization and transport.

It does not handle calling the Elasticsearch or Enterprise Search APIs.

This library uses Faraday by default as the HTTP transport implementation. We test it with Faraday versions 1.x and 2.x.

For optimal performance, use a HTTP library which supports persistent ("keep-alive") connections, such as patron or Typhoeus. Require the library (require 'patron') in your code for Faraday 1.x or the adapter (require 'faraday/patron') for Faraday 2.x, and it will be automatically used.

Currently these libraries are supported:

Use Typhoeus v1.4.0 or up since older versions are not compatible with Faraday 1.0.

You can customize Faraday and implement your own HTTP transport. For detailed information, see the example configurations and more information below.

Features overview:

  • Pluggable logging and tracing
  • Pluggable connection selection strategies (round-robin, random, custom)
  • Pluggable transport implementation, customizable and extendable
  • Pluggable serializer implementation
  • Request retries and dead connections handling
  • Node reloading (based on cluster state) on errors or on demand

Refer to Advanced Configuration to read about more configuration options.


Install the package from Rubygems:

gem install elastic-transport

To use an unreleased version, either add it to your Gemfile for Bundler:

gem 'elastic-transport', git: ''

or install it from a source code checkout:

git clone
cd elastic-transport
bundle install
rake install

Example usageedit

In the simplest form, connect to Elasticsearch running on http://localhost:9200 without any configuration:

require 'elastic/transport'

client =
response = client.perform_request('GET', '_cluster/health')
# => #<Elastic::Transport::Transport::Response:0x007fc5d506ce38 @status=200, @body={ ... } >

Documentation is included as RDoc annotations in the source code and available online at RubyDoc.

Transport implementationsedit

By default, the client uses the Faraday HTTP library as a transport implementation.

The Client auto-detects and uses an adapter for Faraday based on gems loaded in your code, preferring HTTP clients with support for persistent connections. Faraday 2 changed the way adapters are used (read more here). If you’re using Faraday 1.x, you can require the HTTP library. To use the Patron HTTP, for example, require it:

To use the Patron HTTP, for example, require it:

require 'patron'

If you’re using Faraday 2.x, you need to add the corresponding adapter gem to your Gemfile and require it after you require faraday:

# Gemfile
gem 'faraday-patron'

# Code
require 'faraday'
require 'faraday/patron'

Then, create a new client, and the Patron gem will be used as the "driver":

client =

# => Faraday::Adapter::Patron

10.times do
  client.nodes.stats(metric: 'http')['nodes'].values.each do |n|
    puts "#{n['name']} : #{n['http']['total_opened']}"

# => Stiletoo : 24
# => Stiletoo : 24
# => Stiletoo : 24
# => ...

To use a specific adapter for Faraday, pass it as the adapter argument:

client = :net_http_persistent)

# => [Faraday::Adapter::NetHttpPersistent]

If you see this error:

Faraday::Error: :net_http_persistent is not registered on Faraday::Adapter

When you’re using Faraday 2, you need to require the adapter before instantiating the client:

> client = :net_http_persistent)
Faraday::Error: :net_http_persistent is not registered on Faraday::Adapter
> require 'faraday/net_http_persistent'
=> true
> client = :net_http_persistent)
=> #<Elasticsearch::Client:0x00007eff2e7728e0

When using the Elasticsearch or Enterprise Search clients, you can pass the adapter parameter when initializing the clients.

To pass options to the Faraday::Connection constructor, use the transport_options key:

client =
  transport_options: {
    request: { open_timeout: 1 },
    headers: { user_agent:   'MyApp' },
    params:  { :format => 'yaml' },
    ssl:     { verify: false }

To configure the Faraday instance directly, use a block:

require 'patron'

client = 'localhost', port: '9200') do |f|
  f.response :logger
  f.adapter  :patron

You can use any standard Faraday middleware and plugins in the configuration block.

You can also initialize the transport class yourself, and pass it to the client constructor as the transport argument. The Elasticsearch and Enterprise Search clients accept :transport as parameter when initializing a client. So you can pass in a transport you’ve initialized with the following options:

require 'patron'

transport_configuration = lambda do |f|
  f.response :logger
  f.adapter  :patron

transport =
  hosts: [ { host: 'localhost', port: '9200' } ],

# Pass the transport to the client
client = transport)

Instead of passing the transport to the constructor, you can inject it at run time:

# Set up the transport
faraday_configuration = lambda do |f|
  f.instance_variable_set :@ssl, { verify: false }
  f.adapter :excon

faraday_client =
  hosts: [
      host: 'my-protected-host',
      port: '443',
      user: 'USERNAME',
      password: 'PASSWORD',
      scheme: 'https'

# Create a default client
client =

# Inject the transport to the client
client.transport = faraday_client

You can also use a bundled Curb based transport implementation:

require 'curb'
require 'elastic/transport/transport/http/curb'

client = Elastic::Transport::Transport::HTTP::Curb)

# => #<Curl::Easy http://localhost:9200/>

It’s possible to customize the Curb instance by passing a block to the constructor as well (in this case, as an inline block):

transport =
  hosts: [ { host: 'localhost', port: '9200' } ],
  & lambda { |c| c.verbose = true }

client = transport)

You can write your own transport implementation by including the {Elastic::Transport::Transport::Base} module, implementing the required contract, and passing it to the client as the transport_class parameter – or by injecting it directly.

Transport architectureedit

  • Elastic::Transport::Client is composed of Elastic::Transport::Transport.
  • Elastic::Transport::Transport is composed of Elastic::Transport::Transport::Connections, and an instance of logger, tracer, serializer and sniffer.
  • Logger and tracer can be any object conforming to Ruby logging interface, for example, an instance of Logger, log4r, logging, and so on.
  • The Elastic::Transport::Transport::Serializer::Base implementations handle converting data for Elasticsearch (for example, to JSON). You can implement your own serializer.
  • Elastic::Transport::Transport::Sniffer allows to discover nodes in the cluster and use them as connections.
  • Elastic::Transport::Transport::Connections::Collection is composed of Elastic::Transport::Transport::Connections::Connection instances and a selector instance.
  • Elastic::Transport::Transport::Connections::Connection contains the connection attributes such as hostname and port, as well as the concrete persistent "session" connected to a specific node.
  • The Elastic::Transport::Transport::Connections::Selector::Base implementations allow to choose connections from the pool, for example, in a round-robin or random fashion. You can implement your own selector strategy.
  • The Elastic::Transport::Transport::Response object wraps the Elasticsearch JSON response. It provides body, status, and headers methods but you can treat it as a hash and access the keys directly.