Data Observability platform Help

Connections

Connections page is for managing all database connections.

  • Add new connections (by clicking "New connection" button)

  • Manage existing connections (change settings by clicking on a connection)

  • View and manage metadata for a connection, which is used by most DQM functionalities.

  • Check connection status (update status by clicking on status icon)

The list can be filtered and sorted by name, type and connection URL, and shows the synced schema/object/column counts per connection.

Adding a new connection

  1. On Connections page click "New connection" button:

  2. Insert database connection information:

ConnectionsAdd
  • Connector – the type of data source, grouped into databases, files, services (REST, SharePoint, Salesforce) and BI tools. Connectors already enabled are listed first; administrators also see Other (not currently in use) behind the collapsible header at the end of the dropdown, which are enabled automatically on save or test when they need no driver upload (otherwise see chapter Connector settings to add new drivers). Users without the admin role pick from the enabled connectors only. An administrator can also upload the driver right in this dialog; Upload and load connectors restarts the application and brings the dialog back with the chosen connector selected

  • Connection name – custom name for the connection

  • Server – database server address

  • Port – database port

  • Database name – database name in server

  • Properties – additional connection properties

  • User – database user (we recommend using a user with read-only rights)

  • Password – password for the user

  • Parallel queries – maximum parallel connections to database for a user

  • Query timeout (s) – maximum number of seconds the connection waits for a response from the database

  1. Click "Save" button. The "Test" button next to it validates the connection before saving. Test is not offered before the first save for Power BI, BigQuery, Salesforce, SharePoint, Tableau, SAP BusinessObjects and custom connections - for those, save first and test the saved connection.

  2. New connection will appear on the list and connection status will show if connection was successful

Connection certificates

SharePoint and Power BI cloud connections, Salesforce connections and Snowflake connections using key pair authentication sign in with a certificate SelectZero generates. Their form ends with a Certificate section, where Certificate type decides which one the connection uses:

  • Automatic certificate (the default) - the connection gets a certificate of its own when it is saved. Get certificate downloads or copies its public half, and the arrow button next to it regenerates it.

  • Shared certificate - the connection uses a certificate an administrator generated under Settings > Authentication > Connection certificates (see Connection certificates). Upload a shared certificate once where the connections sign in, for example to an Azure app registration or a Salesforce connected app, and every new connection that picks it works without uploading anything new. A shared certificate belongs to no connector or app, so the same one can be used by SharePoint, Power BI, Salesforce and Snowflake connections alike.

Below the choice, the connection shows when its certificate expires. Once a connection is saved on a shared certificate, Get certificate hands out that certificate in the form the connector needs: the public certificate for Azure and Salesforce, the public key for Snowflake (ALTER USER ... SET RSA_PUBLIC_KEY).

  • Only an administrator can change which certificate a connection uses. Other users who can edit the connection see the choice but cannot change it.

  • Moving a connection onto a shared certificate deletes its own certificate. Setting Certificate type back to Automatic certificate gives the connection a new certificate of its own, which has to be uploaded again where the connection signs in.

  • A shared certificate cannot be regenerated from a connection. An administrator renews it under Settings > Authentication > Connection certificates, which names every connection the renewal affects first: they all stop working until the new certificate is uploaded.

  • Switching a connection on a shared certificate to a sign-in without a certificate (an on-premise server, or password authentication for Snowflake) moves it back to Automatic certificate when an administrator saves it.

Generated certificates are valid for one year. Snowflake keeps only the public key, so the date does not matter there.

Adding additional database users

The platform supports configuring additional database users to be used for querying results and seeing results of past executions. The use case for this is if you want to limit some users of the platform to not have access to data that is reachable with the master user configured for the connection.

Additional database users can be configured for existing connections through a menu accessible via the settings button ConnectionsAdditionalUsersButton.png on the right of a connection's master user's username.

Each additional database user will apply for a selection of user roles. The additional database users are ordered, meaning that for querying data, the database user with the highest rank will be chosen that is applicable for the user's roles.

ConnectionsAdditionalUsers.png

In the example above, if a user has the roles Analyst and Developer, then their queries through the platform will be executed using the database user postgres. If another user has roles steward and maintenance, then for them the database user analyst will be used.

If no match is found from the additional users list, then the master user of the connection will be used for executions.

Users with no roles will be skipped when looking for a matching database user.

The database users can be re-ordered by holding and dragging the handle on the left of each row.

Sample connections

Connections example spark

Properties (All are required)

  • transportMode=http - HTTP transport mode for Databricks SQL Warehouse

  • ssl=1 - Enabled encryption over HTTP

  • AuthMech=3 - Specifies authentication mechanism(Personal Access token)

  • EnableArrow=0 - Needed to use traditional row based result handling

  • httpPath=/sql/1.0/warehouses/warehouse-id - Endpoint path for Databricks SQL warehouse

  • ConnCatalog=my_catalog - Specifies driver default catalog for metadata sync

transportMode=http;ssl=1;AuthMech=3;EnableArrow=0;httpPath=/sql/1.0/warehouses/warehouse-id;ConnCatalog=my_catalog;
Connections example redshift
Connections example snowflake

Snowflake connections support two authentication types, selected from the Authentication type dropdown:

  • Password - username + password (default)

  • Key pair - RSA key-pair authentication, recommended for service accounts and required when password authentication is disabled in Snowflake

Setting up key pair authentication:

  1. On the connection form, select Key pair as the authentication type

  2. Enter the Snowflake Username (no password is required)

  3. Click Save - on first save, the platform generates an RSA key pair and stores it encrypted on the server

  4. Re-open the saved connection. Get certificate in the Certificate section at the end of the form now offers three actions:

    • Download - saves the public certificate as a PEM file

    • Copy - copies the PEM-encoded public key to the clipboard, ready to paste into Snowflake

    • Regenerate - generates a new key pair (use when rotating credentials)

  5. In Snowflake, attach the public key to the user with ALTER USER <user> SET RSA_PUBLIC_KEY='<contents>'

  6. Back in the platform, click the connection status icon to verify the connection works

Connection example athena
Example sharepoint
Connections example excel
Connections example csv

Required JAR files:

  • sas.core.jar – core SAS library

  • sas.svc.connection.jar – IOM connection management

  • sas.security.sspi.jar – authentication

  • log4j-api.jar – logging (required by SAS driver)

  • log4j-core.jar – logging (required by SAS driver)

  • glassfish-corba-orb.jar – CORBA ORB (required on Java 11+)

  • glassfish-corba-omgapi.jar – CORBA API

  • glassfish-corba-internal-api.jar – CORBA internals

  • pfl-basic.jar – GlassFish PFL dependency

  • pfl-tf.jar – GlassFish PFL dependency

Optional JAR files (for non-English locales or SAS 9.4M7+ encryption):

  • sas.core.nls.jar – NLS resources for non-English locales

  • sas.svc.connection.nls.jar – NLS resources for connection service

  • sas.nls.collator.jar – SAS-specific collation/sorting

  • sas.rutil.jar – encryption (required for SAS 9.4M7+)

  • sas.rutil.nls.jar – encryption NLS

  • sastpj.rutil.jar – third-party encryption support

Connections example sas
Connections example power bi cloud
Connections example sapbo
Connections example tableau
Connections example salesforce
Connections example rest
Connections example oracle thin

JDBC URL override examples:

jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=TCPS)(HOST=myhost)(PORT=2484))(CONNECT_DATA=(SERVICE_NAME=ORCL)))
Connections example custom

Deleting an existing connection

  1. Click on the connection you would like to delete

  2. Click the "Delete" button in the opened window

  3. Read what the confirmation lists, take a backup if you want one, then confirm

The confirmation does not just ask twice - it counts up everything the delete would touch, so you can see the cost before you pay it.

What the confirmation shows

"Will be permanently deleted" lists each category with its item count: schemas, catalog objects, attributes, descriptions written by your team, custom field values, glossary term links, lineage relations, profilings with their rules and excluded values, profiling schedules, profiling alerts, alerts on catalog objects and schemas, metadata sync schedules, and connection users. Only categories that actually have content are listed, so a short list means there is genuinely little to lose.

"Will be kept, but changed" lists what survives the delete in a different shape: test cases that run against the connection and test cases that compare against it both keep their definitions and lose the connection, reports drop it, and automations whose scope names it are unscheduled. This section starts folded when there is a destructive list above it to read first.

The glossary terms themselves are never deleted - only the links from this connection's attributes to them.

Backup coverage

If you hold the "Allow exports" privilege on the connection and there is something to delete, each destroyed row is also marked in backup or not in backup, and a "Download backup (.zip)" button appears.

The ZIP is composed of the regular exports, so it re-imports through the normal import screens. It carries the connection itself, its objects and attributes, lineage relations, term links, and the profiling objects with their rules, excluded values, schedules and alerts.

Confirming

Whenever anything at all would be deleted or unlinked, you must type the connection name before the "Delete" button becomes active. A connection with nothing to lose - no catalog content and no test cases pointing at it - stays a one-click delete.

Metadata

Metadata is used to automatically create test cases, generate SQLs and manage Profiling. Metadata can be managed by clicking on a specific connection:

Adding Metadata

  • Pick schemas from the schema selector (the refresh button next to it re-reads the available schemas from the database; custom schemas can be typed in and added one at a time):

    AddMetaData.png

  • Click the "Add selected" button to import the picked schemas

  • Metadata should now be available

    connectionsMetadataList
    Additions and changes to objects and columns are synced automatically when the schema has a schedule assigned.

    A schema with no tables yet, such as a new AWS Glue database, can be added too. It is listed with 0 tables and the next sync picks up tables once they appear. 0 tables means the schema is empty, or that the connection's user cannot see its tables: check the user's rights (for Athena, the Glue and Lake Formation permissions) if tables are expected. A schema name the database does not have is refused. File, REST and SharePoint connections still refuse a schema with no content, because for them it means the file or response could not be read.

    If the "Delete" flag is enabled for a schema, then object and column removals from the database are automatically synced.

    Metadata can also be manually updated or deleted by selecting schemas and clicking "Sync selected" or "Delete selected" button. If a sync takes longer, it continues in the background and the pending schemas are marked accordingly.

    "Delete selected" asks for confirmation with the same breakdown as a connection delete, counted across the selected schemas only. Two differences: no backup is offered, because the backup covers a whole connection rather than a subset of its schemas, so nothing is marked as being in a backup; and the schemas themselves stay registered - the delete empties them and clears the sync stamp rather than removing the rows. Because it is a full-schema delete, individual object delete events are not recorded in change history. A schema delete over many objects reports its progress and continues in the background if it runs long.

Disabling objects

Sometimes we need to disable metadata loading for certain objects in the schema to avoid sensitive metadata being usable for all users

  • Hover over the schema which objects you wish to exclude

  • Click on the connectionDisableIcon icon

  • Use the search filter to quickly find specific objects by name

  • Choose all objects you wish to Exclude/Include (You can use Shift+Click to choose multiple checkboxes)

connectionMetadataDisable
NB! When you re-include an excluded object, metadata will be loaded again for given schema

Scheduling Metadata

Metadata can be scheduled to run on a custom schedule on schema level.

metadataScheduleIcons

  • Alerts - subscribe to alerts about the schema's metadata syncs

  • Schedule - assign schedule to schema sync and see last/next sync time. Schemas without a schedule are not synced automatically

  • Delete - Enable if metadata sync should be able to delete missing columns caused by schema drift

Schedule specifics can be found under Schedule

File storage

Users can upload CSV and Excel files directly to application internal storage for easier access. Uploaded files are encrypted at rest.

Uploading files

The connections-file-storage button is found in the Connections page toolbar

connections-files

Files are uploaded/downloaded from the same view. Select the folder to upload into — the "Upload to" hint shows the active target. Hovering a file reveals its download and delete actions.

Several files can be selected at once in the file picker. They are uploaded one after another into the active folder, and the view shows which file is in progress and how far along the selection is, the same way a schema sync does. A file that is rejected, for example because its type is unsupported, it is larger than the upload limit, or it does not pass security scanning, is reported with its reason when the upload finishes. The other files in the same selection are unaffected and still upload.

Creating and editing folder

  • New root folders are created with the "New folder" button; hovering an existing folder reveals a plus icon for creating subfolders. Folders and files can also be dragged to a new location.

  • To edit or delete existing folder hover on folder connections-file-storage-edit-folder and click pencil icon. Deleting a folder asks whether its contents move up a level or are deleted with it.

Using stored files in CSV connection

  1. Set up a new CSV connection

connections-local-csv
  • Connector – Choose CSV connector

  • Connection name – Any custom name for CSV group

  • Storage/Path – Location for CSV files

  • Properties – Default separator for CSV is "," which can be changed based input files

03 November 2026