Emit events on limit reaches

[aaron-seo]

Looking to expand on this issue "Emit event if configured queue length is reached" here as a discussion to gain input from a broader audience :)

We're interested in increasing visibility/operability on some of the limits that have been put in place in the past few releases. As of today, we have some log entries to monitor these limits have been breached (vhost, connection, consumer, queue, etc), but we think it'd also be useful to emit these as events for the event exchange.

Rough sketch:

limit.consumer_max_per_channel.reached

limit.queue_max_length.reached

limit.queue_max_per_cluster.reached

... and maybe for like 90% threshold events, something like:

limit.consumer_max_per_channel.soft_threshold_reached

With this style, users can subscribe to limit.* and what not; but I'm also open to other naming approaches — say consumer.max_per_channel.reached, queue.max_length.reached, and etc

Would appreciate some thoughts and concerns regarding this idea!

[aaron-seo]

Just thinking out loud some of my brainstorming, since the existing limits are already prefixed with the respective resource (i.e. queue.* or consumer.*, even something like consumer.limit.max_per_channel.reached might make more sense, so users can still easily filter in/our the limit events as desired