[RFQ] Listener Cancellation Fix

[trajan0x]

Workaround to be added and documented here prior to launch (#1650)

When an rpc returns an error in listener, the entire listener for that chain shuts off. This of course should not happen because

Would recommend investigating and fixing through unit tests on listener. Scaffolding for this already exists, errors need only be mocked out

[trajan0x]

Investigating this now, you can see the dropoff here in doPoll requests to arbitrum:

The pod breaks at 10:13 and I reset as per @aureliusbtc at 10:23:

As you can see, omnirpc returns an error here:

Because this happens on block.number, we expect an error return here:

To debug this, I'm going to log pollInterval:

Okay, nevermind all that

I think if pollInterval is time.Millisecond (1,000) and we multiply by time.Second we get a lot of minutes. I'm not sure this explained why the thing stalled overnight, but it's worth a shot

[trajan0x]

Fixed