I have a fully serverless environment, and I have a big layer of costs in idle time of GCP Cloud Functions (memory and cpu).

I’ve set the minimum amount of instances to one, and I’m thinking that also this configuration could improve to increase my bill (with minimum 1 instance, GCP will keep always the function warm, and in this case I’m going to pay for it 24/7, pay less, but not zero cost).

Now the question:

1. does gen1 cloud function costs less due to not using cloud run under the hood?
2. does turn the minimum instance to zero, lower my costs(obviously I may have “cold starts”) ?
3. even if I set minimum instance to zero, will GCP keep a little warm my function for small amount of time, or will it kill immediately if no req ?
4. does implement cloud scheduler pre-warm some functions in order to avoid cold starts ?

Thanks

I’ve already tried to read GCP docs, but it’s a little bit confusing