Title: Should there be dns_failure_refresh_rate similar to Cluster.dns_refresh_rate ?

Description:

So the dns_refresh_rate works exactly as expected. But when dns query returns no result, say our deployments are creating one and it takes some time but envoy got deployed before that. Now the DNS do not return any result. In this case envoy retries DNS only after the duration of dns_refresh_rate. Will it be meaningful for envoy to refresh DNS at a different rate when no results are returned ?

[optional Relevant Links:]

https://www.envoyproxy.io/docs/envoy/latest/api-v2/api/v2/cds.proto