Add retry logic around Client#send_request

[jhas3c]

Currently, we don't have any retry logic present if API request fails with some network related exceptions in this line,

response = http.request(request)

we are using Net::HTTP library to make API request and recently it has been seen many request getting end of file reached exception which may be due to this issue.

If we have some retry logic around Client#send_request then it will help to handle these exceptions till some extent.

To add retry logic, we need to define a maximum number of retries, and then wrap the HTTP request in a loop that attempts to resend the request until either it succeeds or it hits the maximum number of retries. We may also want to add a sleep call between retries to prevent quickly exhausting your retry limit in the case of temporary network issues or server-side problems.

module Trolley
  class Client
    #...

    MAX_RETRIES = 3 # Maximum number of retries

    #...
    private

    def send_request(endPoint, method, body = '')
      # your original send_request logic here
      uri = URI.parse(@config.api_base + endPoint)
      #...

      begin
        retries = 0
        response = http.request(request)
      rescue StandardError => e
        if retries < MAX_RETRIES
          retries += 1
          sleep(2**retries) # Exponential back-off
          retry
        else
          raise "Request failed after #{MAX_RETRIES} attempts: #{e.message}"
        end
      end

      if response.code != '200' && response.code != '204'
        throw_status_code_exception(response.message + ' ' + response.body , response.code)
      end
      response.body
    end

    #...
  end
end

Please note that:

1. The MAX_RETRIES constant is defined to specify the maximum number of retry attempts.
2. I've added a begin / rescue block to catch errors that occur during the request.
3. Inside the rescue block, we check if we've hit our maximum number of retries. If not, we increment the retry count, wait for a period of time (sleep(2**retries) provides an exponential back-off), and then retry the request. If we have hit our maximum number of retries, we raise an exception to alert the user that the request has ultimately failed.
4. We can adjust the MAX_RETRIES and sleep values to suit your specific needs.