How rh-aiservices-bu/spark-on-openshift works
This repository is a set of guides, example workloads, image recipes, and OpenShift manifests for running Spark—not an application with its own runtime code. The principal workflow is to choose operator-based or direct submission, run a Spark workload on OpenShift, and optionally retain its event logs in shared object storage for later viewing in the History Server. Separate documented capabilities include custom Spark/PySpark images and Prometheus/Grafana monitoring. Relationships below follow the README; no source excerpts were provided.
Main entry points: examples Spark workloads · README.adoc Usage guide · operator Operator examples · standalone Standalone examples
Components 12
Guidance and examples
- Usage guide
README.adoc - Operator examples
examples/operator - Standalone examples
examples/standalone
Spark workloads
- Spark operator
- Spark workloads
examples
Custom images
- Spark image recipes
spark-images - Spark images
Operations and monitoring
- Object storage
- History Server deployment
spark-history-server/spark-hs-deployment.yaml - Monitoring manifests
spark-operator - Prometheus and Grafana
Other components
- Spark user
Connections 13
- Spark user → Usage guide: chooses workflow
README.adoc - Usage guide → Operator examples: documents operator use
README.adoc - Usage guide → Standalone examples: documents direct submit
README.adoc - Spark user → Spark operator: applies job configuration
README.adoc - Operator examples → Spark operator: configures jobs
README.adoc - Spark user → Spark workloads: submits Spark jobs
README.adoc - Standalone examples → Spark workloads: demonstrates submission
README.adoc - Spark image recipes → Spark images: provides recipes
README.adoc - Spark workloads → Object storage: writes event logs
README.adoc - History Server deployment → Object storage: reads event logs
README.adoc - Monitoring manifests → Prometheus and Grafana: supports monitoring
README.adoc - Prometheus and Grafana → Spark operator: monitors operator
README.adoc - Prometheus and Grafana → Spark workloads: monitors applications
README.adoc