We can start by adding monitoring of minimum API latency (ping or execution of smallest possible requests).