DWH-788: Multi-cluster support

[canbekley]

pr contains

1. implements creation of remote (distributed) tables by providing a remote_clusters section in the profile target such as following:

clickhouse:
  target: test
  outputs:
    test:
      type: clickhouse
      ...
      remote_clusters:
        - name: cluster02
          database_engine: Replicated('/clickhouse/databases/{uuid}', '{shard}', '{replica}')
        - name: cluster03

2. fixes current issues with existing database_engine profile target credential

currently work in following cases:

• when current cluster db engine is Atomic and remote cluster db engine is Atomic
• when current cluster db engine is Replicated and remote cluster db engine is Atomic

not supported by this mr:

• remote cluster db engine is Replicated
• checking existing database engines (if exists, engine is just ignored)

how to test this pr

1. setup environment as described in CONTRIBUTING.MD
2. create the following .env file:

DBT_CH_TEST_USE_DOCKER=True
DBT_CH_TEST_CLUSTER_MODE=True
DBT_CH_TEST_CLUSTER=test_shard
DBT_CH_TEST_REMOTE_CLUSTERS=test_remote_cluster

3. run pytest tests/integration/adapter/remote_clusters/.

[Magicbeanbuyer]

I tried to run the tests locally but they hang for a long time and then all failed...

[canbekley]

@Magicbeanbuyer it is currently failing, because we couldn't figure a clickhouse keeper setup that would allow us to create databases/tables in other clusters. With a common keeper cluster (like we have currently in prod), we run into issues of "table already exists" when trying to create tables with the same name. With a separate keeper cluster (like in this repo test setup), the ON CLUSTER test_remote_cluster ddl's never reach test_remote_cluster, hence get stuck on the initial host.