Currently, it will be fairly silent. We will likely at a minimum be looking to inject logs for developer live logging to indicate the problem and in time we are considering some other options for user facing exception/event notifications longer term. The main symptom one would encounter is that at the rate limit events would be dropped whenever the rate limit is hit. Currently, we use a simple token bucket algorithm for this but we might later add a “penalty box” on top of this which would change the behavior when a rate limit is hit to be a bit different in order to more aggressively prevent buggy drivers from adversely impacting other drivers and the platform.
With the current limits and burst capacity, we don’t expect drivers that are functioning correctly and sanely to bump into the limits. We did have these set too low early on and they were being hit by a developer but in that case things were off by an order of magnitude.
Metrics on this is something we will be looking to monitor during the beta. Fixing that reporting is the actual change included with this firmware release (we were able to adjust the limit parameters based on developer feedback without a firmware release prior to the open beta start).