Memory leak when using RTM

[jacobtomlinson]

We are using this library in opsdroid and have noticed that when leaving the RTM websocket connected for a period of time (24-48 hours) memory will suddenly increase rapidly. We are tracking the issue in opsdroid/opsdroid#743 but I wanted to also raise it here to aid with debugging.

My initial guess would be that the Slack end of the connection is closing or restarting the websocket and there is a bug either here or in the upstream slacker library which is getting stuck and using more memory until it gets OOM killed.

Any assistance with troubleshooting this woudl be much appreciated.