feat(provider): add gcp provider with data

[jonperron]

Changelog

• Update providers.csv
• Add data from gcp with documentation links

Fix #145

[jonperron]

Two remaining questions:

• What "platform" means in this context ? If this is the CPU manufacturers, then Google is sharing a lot of info (cf https://cloud.google.com/compute/docs/cpu-platforms?hl=en)
• GCP has a couple of GPU machine types (https://cloud.google.com/compute/docs/gpus?hl=en), with a range of vcpus and ram. How should they be taken into account ?

[da-ekchajzer]

What "platform" means in this context ?

The platform is the bare metal server that hosts the instance. Since we are calculating the impact of the instance as part of the bare metal server, we need to know its architecture. The platform field must match one of the id's of the available server archetypes. So you should add a platform name that matches a platform you are going to create in boaviztapi/data/archetypes/server.csv. This documentation may help you:

• https://doc.api.boavizta.org/contributing/cloud_instance/#platform
• https://doc.api.boavizta.org/contributing/server/

It is always difficult to get all the data, so sometimes you will have to guess. If so, you should mention it in the warning column. I recommend that you set the CPU name so that the other values are automatically filled in at runtime. If you do this, make sure the CPU is listed in the CPU data file. If it is not listed, you should add it so that everyone can benefit from it: https://doc.api.boavizta.org/contributing/cpu/.

The most important values are :

• CPU.vcpu
• case_type (always rack in cloud contexte)
• CPU.units
• SSD.units
• SSD.capacity
• HDD.units
• HDD.capacity

For POWER_SUPPLY and USAGE feel free to use the values from the other providers.

GCP has a couple of GPU machine types (https://cloud.google.com/compute/docs/gpus?hl=en), with a range of vcpus and ram. How should they be taken into account ?

We believe that GPUs are never shared between instances, so there is no need to include the resources in the instance file. However, you can add the data to the server.csv. For know we are not calculating the impact of GPU but we will be soon so I recommend you to add it.

Hope this is clear, if not ping me on mattermost.