The previous version warned simply if we had metrics set, but that's not precise enough; metrics only really matter if the routes have the same destination. Even then there has to be more than one metric-having route to the same destination for us to actually have a problem, therefore:
First find any routes that specify a route metric, then count them based on the same routing destination. A single route to a destination with a metric is fine (because the metric is meaningless there), but routes which specify the same destination and a metric is not likely to be useful. (There are circumstances where it makes sense, such as a backup/failover
route that's not intended to be preferred)
The previous version warned simply if we had metrics set, but that's not precise enough; metrics only really matter if the routes have the same destination. Even then there has to be more than one metric-having route to the same destination for us to actually have a problem, therefore:
First find any routes that specify a route metric, then count them based on the same routing destination. A single route to a destination with a metric is fine (because the metric is meaningless there), but routes which specify the same destination and a metric is not likely to be useful. (There are circumstances where it makes sense, such as a backup/failover route that's not intended to be preferred)