Add an endpoint_id to the etag caching router to be used in metrics and logs

In #1099 (closed) we've added new metrics that use the caller_id from the context to fill in the endoint_id label.

When verifying this, I noticed more endpoint_ids that were missing, but that did have a feature category set: https://thanos.gitlab.net/classic/graph?g0.range_input=1h&g0.max_source_resolution=0s&g0.expr=sum%20by%20(type%2C%20environment%2C%20endpoint_id%2C%20feature_category)%20(rate(gitlab_sli%3Arails_request_apdex%3Atotal%7Benvironment%3D%27gprd%27%2C%20endpoint_id%3D%27unknown%27%2C%20feature_category!%3D%27unknown%27%7D%5B1m%5D))%20%3E%200&g0.tab=0&g1.range_input=1h&g1.max_source_resolution=0s&g1.expr=count%20by%20(environment)(count%20by%20(environment%2C%20endpoint_id)%20(gitlab_sli%3Arails_request_apdex%3Atotal))&g1.tab=0

These seem to come from etag cache hits, where we add a feature_category to the routes to add it to metrics and logs, we could also add an endpoint_id to the Route, and add it to the context at the same time in the etag-caching middleware.

The fast cache hits will already be included in all metrics and budgets, but it's informational.