I think I’ve worked around it. The web server logs showed some requests coming to the old address so either some relays hadn’t picked up the change or maybe some clients cache the profile information. I’ve now allowed zaps on both the old and new address to my node and I can zap myself on both so is hopefully fixed.