assets import #34573
Replies: 6 comments 5 replies
|
To import a full set of Superset assets (database, datasources, charts, dashboards) via POST /api/v1/assets/import/, your ZIP bundle must have this structure and content:
Example directory tree: Minimal example for each asset (YAML or JSON is accepted): databases/my_database.yaml database_name: my_db
sqlalchemy_uri: postgresql://user:pass@host/db
uuid: 11111111-1111-1111-1111-111111111111
version: 1.0.0datasets/my_dataset.yaml table_name: my_table
database_uuid: 11111111-1111-1111-1111-111111111111
uuid: 22222222-2222-2222-2222-222222222222
columns: []
metrics: []
version: 1.0.0charts/my_chart.yaml slice_name: my_chart
viz_type: big_number
dataset_uuid: 22222222-2222-2222-2222-222222222222
uuid: 33333333-3333-3333-3333-333333333333
params: {}
version: 1.0.0dashboards/my_dashboard.yaml dashboard_title: my_dashboard
uuid: 44444444-4444-4444-4444-444444444444
position:
CHART-1:
meta:
chartId: 1
uuid: 33333333-3333-3333-3333-333333333333
type: CHART
DASHBOARD_VERSION_KEY: v2
metadata: {}
version: 1.0.0Upload the ZIP as the formData field in a multipart/form-data request, with overwrite as a separate string field ("true" or "false") [source]. Common pitfalls: missing dependencies, incorrect UUIDs, or missing/incorrect metadata.yaml will cause import errors. All referenced assets must be present in the bundle, and UUIDs must match across references [source] [source]. For large bundles (over 100MB), use the CLI import tools instead of the API [source]. Let me know if you need a more detailed example for any specific asset! To reply, just mention @dosu. Docs are dead. Just use Dosu.
|
|
@dosu my use case is i have a set having one database connection, 5 datasets from it, 20 charts from these datasets, 6 dashboards from these charts. i have to add all these once for every new tenant registered with only the name of database change in db connection and not disturbing existing resources inside the superset |
|
To onboard a new tenant with their own database, datasets, charts, and dashboards—without disturbing existing Superset resources—you'll need to prepare a unique asset bundle for each tenant. Superset asset import is UUID-based: if a UUID in your bundle matches an existing asset, it will be overwritten (if overwrite=true) or skipped (if overwrite=false). To avoid conflicts, generate new UUIDs for all assets (database, datasets, charts, dashboards) for each tenant and update all references accordingly. There’s no built-in templating or parameterization for changing just the database name; you must update the database connection info and all UUIDs outside Superset before import [source] [source]. Recommended workflow: export your template bundle, script the replacement of the database name and all UUIDs, update all references (datasets to database, charts to datasets, dashboards to charts), then import the bundle with overwrite set to false. This ensures new, isolated resources for each tenant and leaves existing ones untouched [source]. Some teams automate this with scripts or use third-party tools (like the superset-client Ruby gem) to duplicate dashboards and update references for multi-tenant setups [source]. Checklist for each new tenant:
This approach guarantees each tenant gets their own set of resources, and existing Superset assets remain unaffected [source] [source]. To reply, just mention @dosu. Share context across your team and agents. Try Dosu.
|
|
@dosu what about what about slice_id, id in chart yaml's and id, chartId and id fields in dashboard yamls ? |
|
You don't need to include slice_id, id, or chartId fields in your chart or dashboard YAMLs for asset import. These are internal database IDs managed by Superset and are not required (or even used) in the import schema. The import process relies on UUIDs for identifying and linking assets—so just make sure all your UUIDs are unique and all references (like dataset_uuid in charts, or chart UUIDs in dashboard positions) are consistent across your bundle. Superset will assign new internal IDs as needed during import, and will automatically update any chartId references in dashboard layouts to match the new chart IDs it generates. Best practice: omit id, slice_id, and chartId from your import YAMLs and use only UUIDs and reference fields for linking assets. To reply, just mention @dosu. Docs are dead. Just use Dosu.
|
|
@dosu what about in a chart yaml datasource: 1__table, slice_id:1, optionName, dashboards of params and '{"datasource":{"id":1,, "dashboards":[1], "form_data":{"datasource":"1__table","viz_type":"echarts_timeseries_line","slice_id":1 of query context |
|
You don't need to include fields like Best practice: omit To reply, just mention @dosu. Docs are dead. Just use Dosu.
|
|
@dosu slice_name: tenant-1_activepower
|
|
For asset import, you should omit fields like To reply, just mention @dosu. Share context across your team and agents. Try Dosu.
|
|
@dosu 2025-08-12 06:27:30,467:WARNING:superset.commands.database.utils:Error processing catalog 'None' /app/superset/connectors/sqla/models.py:2078: SAWarning: Usage of the 'related attribute set' operation is not currently supported within the execution stage of the flush process. Results may not be consistent. Consider using alternative event listeners or connection-level operations instead. self.database = session.query(Database).filter_by(id=self.database_id).one() /app/superset/connectors/sqla/models.py:2078: SAWarning: Usage of the 'collection append' operation is not currently supported within the execution stage of the flush process. Results may not be consistent. Consider using alternative event listeners or connection-level operations instead. self.database = session.query(Database).filter_by(id=self.database_id).one() /app/superset/models/helpers.py:323: SAWarning: Attribute history events accumulated on 1 previously clean instances within inner-flush event handlers have been reset, and will not result in database updates. Consider using set_committed_value() within inner-flush event handlers to avoid this warning. obj = obj_query.one_or_none() 2025-08-12 06:27:30,731:WARNING:superset.views.error_handling:CommandException Traceback (most recent call last): File "/app/superset/utils/decorators.py", line 259, in wrapped File "/app/superset/commands/importers/v1/assets.py", line 164, in run File "/app/superset/commands/importers/v1/assets.py", line 125, in _import File "/app/superset/commands/dashboard/importers/v1/utils.py", line 69, in update_id_refs File "/app/superset/commands/dashboard/importers/v1/utils.py", line 49, in build_uuid_to_id_map File "/app/superset/commands/dashboard/importers/v1/utils.py", line 50, in KeyError: 'chartId' |
|
The KeyError: 'chartId' happens because your dashboard YAML's position field is missing the chartId key in the meta section for one or more chart nodes. Superset's import logic expects every chart node in the position layout to have both uuid (the chart's UUID) and chartId (an integer, can be a placeholder like 0 or null) in meta. For example: position:
CHART-1:
meta:
chartId: 0
uuid: 33333333-3333-3333-3333-333333333333
type: CHART
DASHBOARD_VERSION_KEY: v2The importer will update chartId to the correct value during import. If chartId is missing, you'll get the KeyError you saw. Just make sure every chart node under position/.../meta/ has both uuid and chartId fields to avoid this error reference. The SQLAlchemy warnings in your logs are not related to this import failure and can be ignored for now. If you want, you can share a snippet of your dashboard YAML and I can review it for you. To reply, just mention @dosu. Share context across your team and agents. Try Dosu.
|
Uh oh!
There was an error while loading. Please reload this page.
POST {{baseUrl}}/api/v1/assets/import/. @dosu please tell me how the zip file or json of bundle input for the given api call should be in-order to set a database connection, datasources from the declared db connection, charts from the declared datasorces and dashboards with charts from declared charts
All reactions