Skip to main content
Version: v2.10.0

Features

This page summarises the features Yeedu provides for two audiences: developers who build and operate Spark workloads, and platform owners who optimize their Spark infrastructure and control cost. The job management features are the universal CLI, logging, dependency management, the Apache History Server and cluster monitoring. The platform features are autoscaling, auto-stop and auto-start, multi-version Spark support, billing breakdowns, notebook support, CUDA support, runtime job arguments, Git integration and AssistantX. Each one links through to the page where you configure it.

Job management features

Universal CLI

We ship a Command Line Interface (CLI) for Windows, Linux, and macOS. You can create, manage, and monitor jobs with a small set of intuitive commands. It's equally useful for interactive work and CI/CD automation. See the CLI commands reference.

Logging

We provide comprehensive logging so you can monitor and debug Spark jobs in real time. Both standard and custom logging levels are available, so you can pick the verbosity that fits your workload.

Dependency management

We detect and install the dependencies your Spark jobs require, and we ship pre-built dependencies for commonly used libraries. You don't have to bootstrap the runtime yourself, so you can focus on the workload instead.

Apache History Server

We run an Apache History Server so you can view and analyze the performance of completed Spark jobs, track progress, inspect detailed metrics, and identify bottlenecks.

Cluster monitoring

We offer detailed cluster monitoring with metrics and alerts for CPU, memory, and network, see Cluster monitoring and Node-level monitoring.

Platform features

Autoscaling

Clusters scale up and down with workload demand, so you only pay for the resources you're actively using. See Load-based sizing.

Auto-stop and auto-start

Yeedu can automatically stop and start clusters on a schedule or when specific conditions are met, so you don't pay for idle time. See Auto suspension and Suspend and resume clusters.

Multi-version Spark support

Yeedu supports multiple versions of Apache Spark, so different teams can pin to the version that suits their workload, and you can upgrade at your own pace.

Billing breakdown

Detailed billing breakdowns let you track infrastructure spend per workspace, cluster, and job, and identify where to optimize. See the usage dashboard.

Notebook support

Notebooks are the primary surface for interactive data science and machine learning on Yeedu. They support multiple languages, collaborative editing, version control, and integrated visualizations. See Notebooks.

CUDA support

Yeedu supports CUDA-enabled NVIDIA GPUs, so you can accelerate ML training and large-scale data processing.

Runtime job arguments

You can launch a job instance with custom configuration and runtime arguments to override defaults for a specific run, without changing the base job definition. See Jobs.

Git integration

Workspaces integrate with Git so you can version-control notebooks and job code alongside the rest of your repository. See Git operations.

AssistantX

AssistantX is the in-product AI assistant that helps you author code, explain results, and debug failures from inside the notebook editor. See AssistantX.