I will test upgrade and test again, though as I stated things seemed to work on the dismiss with that change, so I think we are good there.
I didn't really look into detail as to how you implemented it, I see it is delivering the negative index (last) entry in the event, but should be flexible enough I think.
FYI, I've been running a memory test this morning (with v .1700) where ALL I have done is install the plugin and configure it.
I let it settle after start and it was about at 37,480K. After about 4 hours we are at 50,268. That means we are growing at about 3MB/hour or about 36MB/day.
This is from doing nothing at all...it isn't even receiving events (as i turned them off on my phone).
When it does receive events it could jump by 20-30KB/event.
I don't know enough about the finer points of python, but I expect it is probably something with the websockets. I don't know if this is something you can even fix or if it is an issue with the python libraries.
I'm still monitoring to see if I start getting those crash/hang/not fully working on my production box, it takes a day or two before I see an issue.
Again, I'm currently running to limit to 50MB and it seems to limit it, but I honestly don't know how EG "reclaims" the memory and if doing so is making the issue worse.
So, while I can't say that the memory usage of this plugin is causing my issues - there is clearly a leak.
Honestly, I'd be fine with EG taking up half a GB on this system for what it provides me - but I can't have it flaking out on me.
Right now it is looking like the only solution is a scheduled restart of the PC (or at least EG) nightly until I can figure out what exactly is causing the issue.
Thanks also for the logging info... I will look to set that up... another reason we need a good plugin repo

