I have a Cloud Run service. It uses Redis extensively yet I cannot get request latencies below 100ms:
When I run locally, latencies are sub-10ms. If I check logs, Cloud Run is also reporting sub 10ms latency:
What gives? Cloud Run is behind a load balancer, so my conclusion is that the load balancer is introducing this latency. Any way to debug this or am I just at GCP's mercy? It's a global load balancer, Cloud Run service is in us-central1, I'm in Idaho.