Skip to content

Add a Downstream Connection

Record that the container you are looking at, or one of its fields, feeds another one. The connection you add this way is manual, so no data catalog sync and no Qualytics process will ever change or remove it.

Permission Required

You need Editor permission on every datastore associated with the connection, or the Admin role, to add a lineage connection. See Lineage Permissions for details.

Field reference

The Add downstream panel has two sections, one for each end of the connection. The node you started from is locked into the upstream section, so the downstream section is the one you fill in.

Three tabs at the top decide what you are connecting, and which fields the panel shows. The first is named after what the datastore holds, so it reads Table on a database and File on a file store.

Connects whole containers.

Field Required Type Description
Datastore Select The datastore holding the downstream container. Only datastores where you have Editor permission, including the one you started from. Unrestricted Admins can select any datastore.
Container Select The downstream container itself. Selectable only after a datastore is chosen.

Connects one column to another. Both ends need a field, so the locked section gains a field selector of its own.

Field Required Type Description
Datastore Select The datastore holding the downstream container. Only datastores where you have Editor permission, including the one you started from. Unrestricted Admins can select any datastore.
Container Select The downstream container itself. Selectable only after a datastore is chosen.
Field (downstream) Select The downstream column. Selectable after the container is chosen.
Field (upstream) Select The column on the locked container.

Connects the locked container to a dataset that is not registered in Qualytics, such as a file in object storage, a SaaS application, or a dashboard. You name the dataset the way OpenLineage names it, so lineage your OpenLineage producers later report for the same dataset joins the same node instead of drawing a second one.

Field Required Type Description
Namespace Text The system the downstream dataset lives in, for example s3://reports or snowflake://my-account.
Name Text The dataset's name within its namespace, for example daily/orders.csv.

A dataset that is already registered in Qualytics is refused, with a message to connect its table instead. So is one whose namespace and name point into a datastore where you do not have Editor permission.

Steps

Step 1: Open the container and go to its Lineage tab. From a datastore's Lineage tab, click Open this table's lineage on the container's card instead.

Step 2: Hover the right edge of the node you want to extend, then click Add downstream . You can also right-click the node and choose Add downstream .

Step 3: The Add downstream panel opens. The node you started from is shown locked, with a lock icon, in the upstream section.

Step 4: Choose the mode. Stay on the container tab to connect whole containers, switch to Field to connect one column to another, or switch to External to connect a dataset that is not in Qualytics.

Step 5: In the downstream section, select the Datastore, then the Container. On External, enter the dataset's Namespace and Name instead. See the Field reference above for what each one holds in the mode you picked.

Step 6: In Field mode only, select the Field on the downstream side and the Field on the locked upstream side.

Step 7: Click Add. The button stays disabled until both ends are complete.

Step 8: The new node appears on the downstream side and the canvas re-centers on it.

Tip

Right-clicking a row in a node's expanded field list and choosing Add downstream opens the panel already in Field mode with that field selected, which skips two steps.

Note

When the two fields belong to different containers, Qualytics also creates the container-level connection between those containers if one does not exist yet, so the relationship is visible at both levels.

Note

An external connection keeps pointing at the dataset by name. If that dataset is registered in Qualytics later, the manual connection is not moved to the new container: delete it and add the connection again on the container tab.