We're still seeing this happening. maschmid@redhat.com's testruns show that running tests in parallel hides the issue as multiple reconcile events "unblock" events that might get stuck.

We have a bandaid currently that force reconciles unready ingresses every 30seconds that should in theory unblock this. We don't understand the root cause of it fully though.

We should try to dig down into what causes this and fix it properly to be able to remove the bandaid. Envoy 0.12 seems to work fine as well, but that'll only be released on Maistra 1.1.