Apache Superset™ on the
Stackable Data Platform

Superset connects to SQL databases and query engines via SQLAlchemy. The Stackable image already ships drivers for common backends such as Trino, OpenSearch, PostgreSQL and MySQL; others can be added by installing the driver and configuring a URI. In the SDP, these connections typically point at clusters running alongside as Stackable operators – Trino, for example, is configured as a Database Connection in Superset that points at the platform-managed Trino cluster. Authentication and TLS run through the platform’s secret management.

Superset can be scaled horizontally by adjusting the number of replicas in the Kubernetes Custom Resource Definition (CRD). The operator handles rolling updates and pod lifecycle.

The Stackable operator handles authentication declaratively through an AuthenticationClass: you define LDAP or an OIDC provider (e.g. Keycloak) once as an AuthenticationClass and reference it in the SupersetCluster CRD. Users logging in via OIDC are assigned a default role set through userRegistrationRole. Because the same AuthenticationClass is also used by other SDP components (Trino, Airflow, NiFi …), you configure SSO once, centrally, rather than per product. Superset’s built-in database authentication remains available as a simple starting point; for production, OIDC is the recommended approach.

Yes. Superset supports fine-grained role-based access control (RBAC). Permissions can be assigned to roles, which you then map to users or groups. With Stackable’s operator, these policies can be managed declaratively and version-controlled, which makes it easier to enforce security boundaries across multiple teams.

Yes. The Stackable operator exposes metrics that integrate seamlessly with Prometheus and Grafana. This includes application metrics (e.g. query latency, request counts) as well as container-level metrics. Dashboards for Superset health and usage can be imported into Grafana.

Yes. For restricted environments, you can mirror the required container images and dependencies into a private registry, which the Stackable operator supports configuring. Be aware that additional effort is needed to fetch and update the Python dependencies for connectors.

Yes. Within a single instance, Superset supports multiple teams through role-based separation and separate workspaces. This isn’t strict tenant isolation (everything still runs in the same instance), but RBAC and schema separation in the backend database make multi-team use practical. If you need harder isolation, the Stackable Data Platform makes it just as easy to run several separate Superset instances – each declaratively defined as its own SupersetCluster.

Yes. Superset supports UI customization and theming – colors, logos, CSS – for your own environment. To apply such changes in production, you build a custom Docker image with your modified assets and roll it out via a CI/CD pipeline; the Stackable operator then deploys that image.

Yes. Superset includes an alerts & reports system. You can configure queries or dashboards to be executed on a schedule, and the results delivered via email or Slack. For enterprise setups, you may want to integrate with Airflow to orchestrate refresh and reporting workflows.

Superset can export and import dashboards, charts and datasets, though not all relationships between objects are fully preserved. The more robust approach is a dump of the metadata store (e.g. PostgreSQL), which keeps dashboards, charts, datasets and their relationships intact.

The operator has been tested on major managed and self-hosted Kubernetes platforms: EKS, AKS, GKE, OpenShift, IONOS and K3s. This flexibility lets you deploy NiFi reliably across cloud providers or on-premise infrastructure, while still benefiting from declarative management and operator automation.