The goal is to identify where the bottleneck is (journal-client, svix, webapp, dns, ...?)
The webhook consumer has a strange behavior too: