This website is very useful.
But accessibility through the repository is low because there is no link to the website.
Therefore, it would be better to add a link to the website in the repository details.

When looking for the best family to use you usually care about price per compute unit and not the total price
Once i filter for the minimum instance size i need my next step is usually looking for the instance that cheaper per core/GB

The n2-standard-* machine types show Unavailable in the Linux On Demand and Linux Sustained usage discount cost columns. The pricing is clearly available from Google, so why does it say “Unavailable”?

The following line of code computes C2D as having a 20% sustained use discount. However, C2D does not has sustained use discounts. See:

• https://cloud.google.com/compute/docs/general-purpose-machines#t2d_machines
• "Sustained use discounts apply to only N1, N2, N2D, C2, M1, and M2 machine types." https://cloud.google.com/compute/docs/sustained-use-discounts

I believe this also means this line is wrong for T2A:

What about integrating other currencies?

https://gcpinstances.info/?cost_duration=monthly does not matche https://cloud.google.com/compute/gpus-pricing. And yes of course the google page bills in 1GPU increments.

I'd like pull this information programmatically (to e.g. select a machine type based on user requested vCPU, RAM, etc). Do you have any interest in supporting an API for this info, or would the API just be to scrape it? Thanks for an awesome resource!

The data file has no way to distinguish between x86 vs ARM processor architectures

GCP recently announced an evolution of preemptible VMs. They are now called Spot VMs.

More info here: https://cloud.google.com/compute/docs/instances/spot

Hourly Price in GCP calculator is 37.37$.
Price in GCP website is 42.186$

I need to check with Google support which SKU's are actually in use.

Adding a selected machine type to "selected" in URL fails

<tr role="row" class="odd highlight">
  <td>e2-standard-2</td>
  <td>2 vCPUs</td>
  <td>8 Gib</td>
  <td class="sorting_1">$48.9173 monthly</td>
  <td>$48.9173 monthly</td>
</tr>

Hello !

Thanks for your work !

When I set the Currency to Eur, ordering of items changes, and I can't sort by machine type , or SUD. But sorting by cpu and memory works as normal.

I'm on Chrome on macOS.

Thanks !

Benchmark data is missing for several instance types, e.g.:

• a2 family
• m3 family
• c3 family
• n2-highcpu-96
• n2d-highcpu-128
• n2d-highcpu-224

Intro blog post: https://cloud.google.com/blog/products/compute/compute-and-infrastructure-enhancements-at-next24

Pricing details: https://cloud.google.com/compute/vm-instance-pricing

Just checked, no data at all are imported. No idea what's up.

For example N2Ds are shown in us-west2:

(source: https://gcpinstances.doit-intl.com/?min_memory=30&region=us-west2&cost_duration=monthly)

...while according to the GCP docs N2Ds are not available in this region.

This also applies to C2Ds shown on this screenshot and probably other types on other regions.

It would be great to add the feature to enable inter region comparisons.

Thank you for your project! It's being very useful to me.

One feature that I missed is the ability to get prices of instances using MS SQL.

scraper.py file is ready, waiting for GCP to update the calculator.

The list of instances isn't complete, even for machine types that have been around for some time. For example, n2-highmem-96 is missing while all other n2-highmem machines are present.

Hi @Burekasim and the whole team!

Thanks again for this solution. While there are alternatives now, I really like your minimalistic and clear site.

But again I don't feel confident that it shows the up-to-date info because as of writing these words the data update time is 2022-12-18, which is 2.5 weeks ago if you are an optimist ...or last year if you are a pessimist. 😅

Anyway, I wanted to ask what is the progress on automating the updates? I read the comment that's it's on the roadmap from June 2022 but I also remember my ticket from October 2021 about the same thing, so perhaps I can help you with this if you are stuck?

Don't want to brag but I have a lot of experience with various CI/CD solutions, including Github Actions which would make the most sense here probably. (You can check out my work in gitlabform/gitlabform, for example.) I propose to create a site update pipeline running once per day (or on another schedule, if you think that makes more sense).

What do you think?

The e2-medium vCPU count shown as 1. I have it deployed and I see 2 vCPUs. Official docs also show it as 2 vCPU instance.

I have recently written an internal app that is using Infracost API and it works really well.

Infracost API is a third-party solution, not from GCP, but from a company that focuses on providing accurate cloud infra cost estimations. So they do their best to deal with the price scraping / using Catalog API in the backend and provide a nice GraphQL API in the front.

It's available for free but you can also self-host your own instance.

Google provides benchmark data in the form of Coremark scores, so to evaluate price/performance, it would be nice if this site also scraped those scores in a way that allowed things to be correlated, either by adding them as columns to the existing sheet or using a key that we could join against.

Linux benchmarks: https://cloud.google.com/compute/docs/benchmarks-linux
Windows benchmarks: https://cloud.google.com/compute/docs/benchmarks-windows

Hi!

GCP docs provide info about E2 shared-core (e2-micro, e2-small, e2-medium) network badwidth limits, see https://cloud.google.com/compute/docs/general-purpose-machines#e2_machine_types but your app shows that it is "Unknown".

Is it a bug/missing feature or is there a reason for that?

Hi, thanks for a great tool!

It would be even greater if it would be possible to filter numerical columns as range.

This would allow queries like: "show me the instance types that provide >= 60GB of RAM".

I think that this could make it possible pretty easily: https://datatables.net/plug-ins/filtering/row-based/range_numbers

https://gcpinstances.doit.com seems to be down. 😢

Hi,

There are some inconsistencies that make loading the data file difficult. In particular:

1. At the top level, there is a map keyed by machine family...except that it also includes licenses, which has a completely different data definition
2. Within a machine type spec element, the cores element is sometimes an integer value (e.g. n2d-highcpu-16 == 16) while other times, it is the string value "shared" (e.g. f1-micro)

Would there be any appetite here to make the data format more standardized? I'm thinking we could start with a json-schema definition.

Need to update T2D specifications in the scraper.py script, I don't know if GPU/Sole tenant/etc is available (there is no documentation):

'gpu': 0, 'sole_tenant': 0, 'nested_virtualization': 0, 'regional_disk': 0

Hi,

I think that since the style has been updated in #38 (and I agree that the site is prettier now!) the dropdowns "Region" and "Columns" don't scroll.

This is a problem especially for the Regions as the list is long enough to not fit on my big monitor and I have to resort to scrolling with keyboard arrows.

Hi @Burekasim,

I just noticed the CPU Type info is out of date. For example, gcpinstances.info shows most n2-* machine types as having only Cascade Lake CPUs available, while in fact it supports both Cascade Lake and the newer Ice Lake.

I'm not sure if you're able to pull this data from an API, but in case it's manually entered I've gone ahead and taken screenshots of the all the CPU types available for each machine type in case it's helpful for you.

Note: in the first few screenshots you'll notice a CPU Type that is greyed-out. That's not actually a CPU type available on that machine type, it's just displayed because it was the CPU type that was selected before I changed the Machine type. Only the non-grey CPU types are available for the machine type.

Hi,

I noticed that https://gcpinstances.info 's footer says:

How? Data is scraped from multiple pages on the GCP site. This was last done at 2021-08-23 06:30:00 UTC.

That's a long time ago. Can you please either start updating the site again or take it down, if you don't plan to? Otherwise it may mislead people.

The idea is to show instance price for all regions at a single view.

Use case example : I can run my QA on any region. I need to find the cheapest region for running n1-highcpu

Such as here https://www.instance-pricing.com/provider=gcp/instance=n1-highcpu-2

On the current solution I need to check the price per region for all regions. Very time consuming.

Hello,

It would be handy to "keep" the choices in the list to ensure it can be bookmarked with the same ones.
For exemple, when you choose a region, cost or currency - and you bookmark, then you reload the webpage: all the choices are reset to default.

Thank you,
Br,
François

Thanks for the site.
The region is reset for every request and region variable doen't work

All non-USD currencies are showing 0. monthly in the table, eg:

All G2 instances are missing. Details can be found on this page: https://cloud.google.com/compute/docs/accelerator-optimized-machines#g2-vms