Add support to enable the use of different services for building models

**Is your feature request related to a problem? Please describe.**
Some users want to have specific compute-compute separation so that certain models can be executed within specific services (ClickHouse clusters).

**Describe the solution you'd like**
Solutions similar to those provided by Snowflake and Databricks would make a lot of sense:
- Databricks: https://docs.getdbt.com/reference/resource-configs/databricks-configs#selecting-compute-per-model
- Snowflake: https://docs.getdbt.com/reference/resource-configs/snowflake-configs#configuring-virtual-warehouses

**Describe alternatives you've considered**
No alternatives at this time, apart from building parts of the projects with different `profiles.yml` files.

**Additional context**
We need to test this thoroughly, especially regarding what happens if the target cluster is idle at the time of execution.

Start by reviewing the profiles.yml approach mentioned in the issue and the linked dbt configuration examples for Databricks and Snowflake. Define the per-model service configuration and test behavior when the selected ClickHouse cluster is idle; the issue is complete when model execution can target different services without separate profiles.yml files.

Written by the indexing model from the issue text.