Larger VM sizes for machine executor?

[ktham]

It currently is spec’ed at 2CPU, and 8GB RAM. Is there an option to choose a larger VM size? We’re not able to use docker executor given that we also leverage docker/docker-compose inside our tests, and we’re looking to find ways to improve our build times, currently, our workflow takes over an hour to complete, whereas on our macbook, it takes about 8-10 minutes.

One big bottleneck is Scala compilation for our monolithic Scala app, which is really slow on 2CPU.

Any help appreciated, thank you!

[halfer]

I’m not aware that Machine boxes come in different sizes. However, these thoughts comes to mind:

• I use Docker Compose on Docker, and it works very well, with a minor problem to do with volumes. If you can go down this route then an 8 CPU 16G monster is available for an additional fee.
• If you are sure DC won’t work on Docker for your situation, then you can split your compilation into a separate job, set that as the monster Docker executor, and then make that available to the DC job, which runs in Machine, and then join the jobs together in a Workflow.

[ktham]

Thanks for your suggestion, but the way we use docker-compose (which we use for our database containers) we use port mappings, to map containers ports to localhost, and docker commands in a docker executor run in a remote environment, so I don’t think it will work for our use-case.

We need bigger VM sizes for machine executors. This is a very simple change and we are willing to pay more for this.

[halfer]

@ktham

We need bigger VM sizes for machine executors. This is a very simple change

This may just be a non-standard request that CircleCI are not willing to support; I don’t recall it being in the docs. At this point, I’d recommend opening a support ticket to ask. Let us know how you get on!