class Azure::Core::Http::RetryPolicy

A HttpFilter implementation that handles retrying based on a specific policy when HTTP layer errors occur

Attributes

retry_data[RW]

Public Class Methods

new(&block) click to toggle source
# File lib/azure/core/http/retry_policy.rb, line 25
def initialize(&block)
  @block = block
end

Public Instance Methods

call(req, _next) click to toggle source

Overrides the base class implementation of call to implement a retry loop that uses should_retry? to determine when to break the loop

req - HttpRequest. The HTTP request _next - HttpFilter. The next filter in the pipeline

# File lib/azure/core/http/retry_policy.rb, line 37
def call(req, _next)
  retry_data = {}
  response = nil
  begin
    response = _next.call
  rescue
    retry_data[:error] = $!
  end while should_retry?(response, retry_data)
  if retry_data.has_key?(:error)
    raise retry_data[:error]
  else
    response
  end
end
should_retry?(response, retry_data) click to toggle source

Determines if the HTTP request should continue retrying

response - HttpResponse. The response from the active request #retry_data - Hash. Stores stateful retry data

The #retry_data is a Hash which can be used to store stateful data about the request execution context (such as an incrementing counter, timestamp, etc). The #retry_data object will be the same instance throughout the lifetime of the request.

If an inline block was passed to the constructor, that block will be used here and should return true to retry the job, or false to stop exit. If an inline block was not passed to the constructor the method returns false.

Alternatively, a subclass could override this method.

# File lib/azure/core/http/retry_policy.rb, line 68
def should_retry?(response, retry_data)
  @block ? @block.call(response, retry_data) : false
end