slack-ruby-client: RTM connection fails in large org but works in small org

Hello,

I’m attempting to use a bot to connect Slack with Pagerduty: https://github.com/stripe-contrib/pagerbot

• I set up the bot with a Slack bot access token.
• I verified that the token works by using “curl” from the command line and was able to see the message getting posted in Slack.
• I tested the bot on a large org (5000+ people) and it consistently fails to connect.
• I tested the bot on a small org (15 people) (using a different token of course), and it worked fine.

I’ve worked with the developer of the bot @macobo, and he suggested I file an issue here.

I’ve created a gist with the error logs from heroku. The first log shows the errors when running this in the large org. The second log shows the successful connection with a smaller org:

I added some comments at the bottom noting some differences, notably:

1. On large org I see this in the response, but not in the small org:

transfer-encoding: "chunked" 

Could there be an issue with the large number of members in the org or the number of channels causing something to stall? Is there a huge payload coming back from the RTM API due to my large org?

1. There is a delay of 100+ seconds in the large org log, and the error:

Socket URL has expired

Thanks for looking!