Dear Pulumi experts, I hope this message finds yo...
# general
r
Dear Pulumi experts, I hope this message finds you well. I have implemented an infrastructure as code (IaC) solution using Pulumi (love it, keep on doing the good work), and I would like to ensure that it has been implemented correctly. Specifically, If you have such metrics, I would like to understand how long it should take to deploy a certain number of resources. Could you please provide me with some insights regarding the expected time for deploying a specific number of resources using
Pulumi
? For example, if I were to deploy 1000 resources, approximately how long should it take? Your expertise and guidance in this matter would be greatly appreciated. Thank you in advance for your assistance. Best regards, Moshe.
s
I’m afraid there are too many variables involved to be able to provide a rough estimate of how long it would take to deploy 1000 resources (what kinds of resources, what sorts of dependencies, latency to the nearest cloud provider API endpoint, which cloud provider, etc.). Is there a reason you are asking?
r
I have approximately 1000 resources in my state, involving almost all kind of resources with several dependencies. The reason I'm asking is because I wanted to know if it's normal to wait 12 minutes for each deployment. Moreover , I running several times 'pulumi up' with out any changes and yet each run is about 12 minutes. I was wondering, in average what are the numbers for such case? Do you have any metrics that could help ? Thanks 🙏
s
There really are too many variables to be able to provide an estimate. For example, deploying a managed Kubernetes cluster (EKS/AKS/GKE) takes anywhere from 6 minutes to 15 minutes. That’s an example of (essentially) the same resource being provisioned on different cloud providers, and you can see the difference in deployment times.
r
First , I'm not using any kubernetes cluster. What do you say about the fact that I'm running same code with any diff and it takes that much time, is that make sense? I was thinking that if state is the same and there is not to update in the cloud it should take less..
s
I understand you’re not using Kubernetes; that was merely an example to show that there is quite a bit of variability in deployment times across different cloud providers, even for (essentially) the same resource.
I’m running same code with any diff and it takes that much time, is that make sense
If Pulumi is performing a refresh (where it queries the cloud providers for the state of the resources), then the operation is subject to whatever delays or time requirements imposed by the cloud providers’ APIs. You can skip the refresh and allow Pulumi to work from its saved state, if you’d prefer, and that would probably speed things up. (The tradeoff is that the saved state may be out of date with the cloud resources because no refresh was performed.)
r
Thanks again, Actually this is what I'm currently doing, Running update without refresh. (I mentioned k8s since i thought it's one of the big deployments.. nevermind... 🙏 )
s
Gotcha. If you’re concerned that something is wrong, or taking excessively too long, I’d recommend filing a GitHub issue with more details, and we can troubleshoot more extensively/in-depth from there.