Hey everyone, I am using meltano with an sqlite state backend and a configuration that looks like this:

version: 1
default_environment: dev
project_id: 0199dcf7-566b-717d-bf2d-eaa09b89fa49
environments:
  - name: dev
    config:
      plugins:
        loaders:
          - name: target-postgres
            config:
              database: postgres
              host: localhost
              user: postgres
plugins:
  extractors:
    - name: tap-facebook
      variant: meltanolabs
      pip_url: git+<https://github.com/MeltanoLabs/tap-facebook.git>
      config:
        start_date: 2025-01-01
      select:
        - campaigns.*
        - adsets.*
        - ads.*
        - adsinsights_default.*
        - videos.*
        - images.*
        - creatives.*
  loaders:
    - name: target-postgres
      variant: meltanolabs
      pip_url: meltanolabs-target-postgres
      config:
        add_record_metadata: true
jobs:
  - name: meta-ads-to-postgres
    tasks:
      - tap-facebook target-postgres
schedules:
  - name: business_1
    interval: "@daily"
    env:
      TARGET_POSTGRES_DEFAULT_TARGET_SCHEMA: business_1_data
      TAP_FACEBOOK_ACCOUNT_ID: "XXXX"
    job: meta-ads-to-postgres
  - name: business_2
    interval: "@daily"
    env:
      TARGET_POSTGRES_DEFAULT_TARGET_SCHEMA: business_2_data
      TAP_FACEBOOK_ACCOUNT_ID: "XXXX"
    job: meta-ads-to-postgres

So basically I want to sync Facebook ad data from 2 separate businesses to 2 separate postgres schemas. So after running sync for 1 business, I checked the state table in sqlite and saw this:

sqlite> select * from state;
prod:tap-facebook-to-target-postgres|{}|{"singer_state": {"bookmarks": {"adimages": {"replication_key": "updated_time", "replication_key_value": "2025-10-16T18:04:44.006000+00:00"}, "ads": {"replication_key": "updated_time", "replication_key_value": "2025-10-17T00:24:48+1100"}, "adsets": {"replication_key": "updated_time", "replication_key_value": "2025-10-16T09:56:49+1100"}, "adsinsights_default": {"replication_key": "date_start", "replication_key_value": "2025-10-17"}, "advideos": {"replication_key": "id", "replication_key_value": "999998314788384"}, "campaigns": {"replication_key": "updated_time", "replication_key_value": "2025-10-10T00:08:44+1100"}, "creatives": {"replication_key": "id", "replication_key_value": "998775395547354"}}}}|2025-10-16 18:10:15.088932

Basically I see that the state for incremental replication is stored on the extractor level. Since each schedule uses the same extractor but gets the data from a different ad account would't that break the incremental replication for the second sync? Let me know if I am missing something and I would appreciate any feedback if this is not the best design pattern for the meltano config. Still a meltano newbie. Thanks in advance!

If you add an extra inherited

tap-facebook

you can give them different namespaces to avoid state collision https://docs.meltano.com/concepts/plugins/#plugin-inheritance

So something like this?

version: 1
default_environment: dev
project_id: 0199dcf7-566b-717d-bf2d-eaa09b89fa49
environments:
  - name: dev
    config:
      plugins:
        loaders:
          - name: target-postgres
            config:
              database: postgres
              host: localhost
              user: postgres
plugins:
  extractors:
    - name: tap-facebook
      variant: meltanolabs
      pip_url: git+<https://github.com/MeltanoLabs/tap-facebook.git>
      config:
        start_date: 2025-01-01
      select:
        - campaigns.*
        - adsets.*
        - ads.*
        - adsinsights_default.*
        - videos.*
        - images.*
        - creatives.*
    - name: business1_ extractor
      inherit_from: tap-facebook
      config:
        account_id: "XXXX"
    - name: business2_extractor
      inherit_from: tap-facebook
      config:
        account_id: "XXXX"
  loaders:
    - name: target-postgres
      variant: meltanolabs
      pip_url: meltanolabs-target-postgres
      config:
        add_record_metadata: true
    - name: business1_loader
      inherit_from: target-postgres
      config:
        default_target_schema: business1_data
    - name: business2_loader
      inherit_from: target-postgres
      config:
        default_target_schema: business2_data
jobs:
  - name: meta-ads-to-postgres-job
    tasks:
      - business1_extractor business1_loader
      - business2_extractor business2_loader
schedules:
  - name: meta-ads-to-postgres-schedule
    interval: "@daily"
    job: meta-ads-to-postgres-job

Yes that should work. Assuming you want complete separation of your output data to different schema. You might be able to retain the same target schema as account id is used as a key in many of the output tables. Depends on how concerned you are to keep the data completely separate.

Currently I want to keep the data separate so this works