Looking Glass — How the Internet Routes to You From Elsewhere
A looking glass is a public interface onto another network's routing table. Querying one shows how that network reaches a given prefix: the AS path, the next hop, and which announcement it prefers. It is the only practical way to see your own routing as others see it, rather than as you announce it.
Why run it
Because announcing a route and having it accepted are separate things. You can see your announcement leaving; whether it propagated, whether it was filtered, and whether a more specific announcement from somewhere else is being preferred are all questions only an external vantage point answers. When one region cannot reach you and everywhere else can, querying a looking glass in that region is usually the fastest route to an explanation.
Why automate it
Routing changes propagate globally and silently. A peering relationship lapses, a transit provider adjusts a policy, an upstream applies a filter — your announcement is unchanged and your reachability is not. Sampling several vantage points on a schedule turns that into something observable, and it detects the serious case as a side effect: an unexpected origin AS for your prefix means someone else is announcing your address space, which is a hijack and needs to be known about in minutes rather than whenever a customer reports it.
What it gives you
An outside view of your own reachability, which is the only view that matters and the one you cannot get from your own routers. Monitored, it also provides hijack detection for the prefixes you originate.