To better observe how the rate limiter is affecting individual clusters / nodes, we need a dashboard showing establishment metrics (current rates, queues, rejections, etc) for each node in a cluster, alongside various generic health metrics for those nodes.

This will be used both to determine effectiveness of the static limits (e.g. if it is too aggressive and it’s limiting establishments when the system has many available resources, or if it's too conservative and it kicks in later than it should have) and for incident response.