[Webinterface] After mainnet, revisit polling

[rfornea]

There are probably many solutions to this problem. Long term there may be iota events we can listen for, but that's a ways away.

But currently, requests to the iota node can "pile up" before any of them complete. My suggestion is wait for polling requests to complete before issuing subsequent requests. It would be better if we would wait for each poll to come back, before polling for the same iota addresses again. Basically still poll, but just not on a strict interval, and don't issue subsequent GETs to iota node until the previous has resolved.

Discuss other options in the comments.

[AaronVasquez]

Is the problem that we don't wait for a response before polling again? Or are we just making too many requests in general?

[rfornea]

I think it is likely a combination of both. Even with a decrease in the amount of polling they still seem to be bogging down.