Compute Engine

Welcome to the Google Compute Engine idea forum. You can submit and vote on ideas here to tell the Google Compute Engine team which features you’d like to see.

This forum is for feature suggestions, and is visited regularly by Google’s Product Managers. It is not actively monitored by our technical support team. If you’re looking for help forums, look here:

If you are looking to file a bug report for which you do not need 1:1 technical support, you can also visit our Public Issue Tracker

We can’t wait to hear from you!

  1. Need to route from forwarding rule port to another port for instances in target pool

    Not sure if this is already supported, but my scenario is that I'd like my load balancer to listen for TCP traffic on port 80 and forward traffic to instances in the relevant target pool listening on port 8080.

    I tried to find documentation support the above use case to no avail.

    20 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    under review  ·  3 comments  ·  Flag idea as inappropriate…  ·  Admin →
  2. Firewall should (optionally) allow IP protocol 41

    Adding firewall support for IP protocol 41 (IPv6 encapsulation in IPv4) would allow "sort-of-IPv6" communication without as much work as full IPv6 support.

    Issue #8 (IPv6 support) is probably much harder; I suggest protocol-41 support could be a good halfway house.

    If native IPv6 support can be added, this would not be needed. (Although other people might want per-protocol control in the firewall for other reasons e.g. SCTP?)

    I think it would be perfectly reasonable to have a simple yes/no rule for non-TCP/UDP protocols; I presume anyone turning them on would be able to apply any needed fine-grained control at…

    60 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    under review  ·  2 comments  ·  Flag idea as inappropriate…  ·  Admin →
  3. Create multiple vms at once.

    I would like to create multiple instances with the same template.

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    You can do this with Instance Templates and Managed Instance Groups by creating an Instance Template and then creating a “MIG” that uses the Template, and setting it’s target size to the number of VMs you’d like.

    If this doesn’t accomplish what you need, please post back comments or edit your request to provide more details on your scenario and what you’d like to see, and we’ll be happy to consider your feedback.

  4. allow instance templates to be based on image family instead of actual image

    In the instance template config from the Console there should be an option to set the image to a family instead of an actual image. This is possible with the API though.

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  5. Show imports in code snippets

    Code snippets are much harder to actually leverage when the critical piece--the import headers--are omitted.

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  6. completely reset

    Hi.

    Good service, but i dont see completely reset instance button to default settings, because i see previous installed version of PHP etc...

    2 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Sorry, but it’s difficult to understand the scenario you’re describing. Please provide more detail (edit your suggestion or add comments) about how you created your instance, what settings you’re looking to reset, etc. so we can consider your feedback. Thanks!

  7. Provide autoscaling but with support for public IP address for each new VM that is pre-configured and newly spawned.

    Sometimes there are institutions that need a public IP on the VMs to be whitelisted before they allow access to their websites/services.

    6 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    2 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Hi, could you provide an example or more detailed description of the scenario you’re describing? It sounds like your best approach might be to allocate static IPs, so that you know in advance of creating the VM what the IP is going to be (and can keep it as long as you want). Then assign those IPs to the appropriate VMs once you have configured the remote networks. Does this address your concern?

  8. I have applied for nvidia tesla k80 GPU but i can't see the option to monitor it

    I really appreciate when you came up with the feature of Nvidia tesla K80 GPU but in addition you can add feature for GPU monitoring.I can't see the feature of monitoring activities of GPU neither in windows server 2016 nor in google console.Please enable the gpu in case if you have disabled in windows 2016 server us-east -1c.Thank you

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  9. A proper way to autoscale MIGs using memory usage

    The autoscaler for a Managed Instance group does not allow for a memory usage based autoscaling policy from Stackdriver metrics which actually meets real world needs. I have raised in the Google Groups community, so just linking to that instead of retyping - https://groups.google.com/forum/#!topic/gce-discussion/y8vWBaJFflA

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    under review  ·  0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  10. 1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Thanks for your request. We currently do not offer a cloud region that is physically in Chicago. Our general philosophy on data centers is to choose locations that serve large populations well, but which are also located in places that lead to affordable pricing. As a result, one of our regions in the Midwest that many customers in Chicago leverage is us-central1, which is based in Iowa. Whether or not this works for you will depend on your network connectivity needs, of course.

  11. Make possible to hibernate a VM

    It would be nice to hibernate (save status and RAM) of a running instance and store it as snapshot.

    6 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  12. Processor usage chart on micro and small instances.

    Processor usage chart on micro and small instances is unclear. It should be clear if 100% on the graph indicates the available power for the shared instance or the entire processor.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  13. Method regionBackendServices.getHealth does not work as documented

    Calling regionBackendServices.getHealth method documented here:

    https://cloud.google.com/compute/docs/reference/rest/v1/regionBackendServices/getHealth

    Returns the following error:

    Invalid value for field 'resource': ''. Specified instance group is not listed in the backend service.

    The documentation does not mention a field named "resource" and there appears to be no way to pass a value for this field. Values passed in the body as resourceGroup appear to be ignored.

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Thanks, I’ve asked a responsible team member to take a look. Note, this forum is only for feature requests, so for better responsiveness in the future when reporting bugs/issues please visit the gce-discussion Google Group, or file an issue report on our Public Issue Tracker.

  14. The Free Tier description might not be very clear

    I would suggest to use the same naming in the Free Tier guide with respect to console naming.
    For example
    "1 f1-micro instance per month (US regions only - Excluding Northern Virginia)" is not clear if is northamerica-northeast or us-east.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  15. gcloud beta compute instances create-with-container multiple network interfaces support

    gcloud beta compute instances create-with-container command should support creation of multiple network interfaces same as plain create command.
    currently it ignores the definition of other --network-interface beyond the first one.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Hi, we believe this is a bug in gcloud. Similar functionality is possible through the Console UI (website). Our team is aware of the issue, however you may find it more helpful to file a bug in our Public Issue Tracker. To avoid a goose chase, I’m leaving this entry open here for now, but please be aware that UV is mostly for Feature Requests and staffed lightly, unlike the Issue Tracker which is staffed by our support team (and integrated into our internal bug tracking). If you’d like to file an Issue, the GCE tracker 1 please go ahead, and 2 has more information on how to use our Issue Trackers in general. Thx!

    1 https://issuetracker.google.com/savedsearches/559755
    2 https://cloud.google.com/support/docs/issue-trackers

  16. Can't find options to set up GPU.

    I came here to try out Compute Engine with GPUs. Blog Posts told me that this is possible. I thought I upgraded to a paid account and added my credit card. However I can't find the GPUs listed under "quota".

    I suspect that your philosophy is that serious users will be able to navigate this byzantine system, but it is frustrating not to be able to get started; in fact that prevents me from BECOMING a serious user.

    Hmm, how can you tell if your account is “upgraded” so that you should be able to see the “Google Compute Engine”…

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Sorry to hear about your experience, we’ll look into making the experience easier to use. In the meantime however, please visit the gce-discussion Google Group if you need some tips in the right direction to get started.

  17. region api needs to include the location information

    I would like to have information about the location of a region included in the compute#region kind.

    I would like to be able to use the region/list api (https://cloud.google.com/compute/docs/reference/rest/v1/regions/list ) to query for regions that are located in a specific country.

    But since the location information is not available this is not possible.

    Only alternative at this moment is to create this api myself and maintain the information manually based on the https://cloud.google.com/compute/docs/regions-zones/ page that includes the location in the "Available regions & zones" table.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  18. Edit VM instance no price shown

    When I edit a stopped VM instance and change CPU count then no price estimate is shown. Expected behavior: Price estimate is shown like when creating a new VM instance.

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  19. Add GPU utilization graphs

    This would be similar to the CPU utilization graphs that are associated with each vm instance.

    9 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
    under review  ·  Paul Nash responded

    Thanks, we’ll consider this in the future. Because our GPUs are a “pass through” operational mode, it’s not clear that GCE has access to the metrics that would be necessary to implement this.

  20. Sharing custom images based on service account and not user email ID

    Currently, if we want to provide a user the permission to create instances using images belonging to another project, the user's email ID needs to be added to the project and granted the compute.imageUser role.
    Could this be done based on the service account ID and not user email?

    For example project A has a user 'abc@gmail.com' who wants to create instances in project A using images in project B. Currently we would have to add 'abc@gmail.com' as a compute.imageUser in project B.

    It would be nice to do it by registering [PROJECT_NUMBER]@cloudservices.gserviceaccount.com as the compute.imageUser…

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  • Don't see your idea?

Compute Engine

Categories

Feedback and Knowledge Base