Recently CI asan jobs started flaking. Log inspection shows that it looks like jobs are getting killed,
suggesting we're short on memory. I suspect that this may have started since the last Envoy update, but I am not sure.
This needs looking into.