I'm running a sample Google Cloud Function which is triggered by a Pub/Sub subscription. Function is CPU bound (does calculations which take 100-300ms per call).
I've sent 1 million messages to the topic. The function ramps up quickly but plateaued at about 800 calls/sec and never goes above 1000 calls/sec. So, it took about 20 minutes to process all messages.
What is the way to diagnose why this rate limit occurred? Is there a chart or metric which would show which quota I hit?