Cloud Run meaning fo the metric container startup_latencies

Google Cloud Serverless Applications
1

I’m doing some work on performance monitoring and have been investigating instances startup times. I see that GCP has the container startup_latencies metric.

The description says: Distribution of time spent starting a new container instance in milliseconds.

Does this means that this is the time from starting the container until it is ready to receive request? Does this time is before or after startup probes come in (if they’re configured)?

2

Hello @lguida ,

Welcome to Google Cloud Community!

startup_latencies metric is still in beta based on this Cloud Run reference. This basically means that the Cloud Run instance is being measured from the time of its deployment in milliseconds. This metric is also being sampled every 60 seconds and after sampling, data would not be visible for up to 180 seconds.

Hope this helps.