ElasticJob is a lightweight, decentralized distributed scheduling service from the Apache ShardingSphere project, built for Java developers and operators who need sharded, highly available scheduled jobs rather than a single-node cron daemon.

What it is

ElasticJob is a distributed scheduled job framework written in Java and released under the Apache License 2.0. It became an Apache ShardingSphere sub-project on May 28 2020. The project combines flexible scheduling, resource management and job governance behind a unified job API, so that developers write a job once and deploy it wherever it is needed. It is a lightweight, decentralized solution that provides distributed task sharding services, and the feature groups its README names are Elastic Schedule, Resource Assign, Job Governance and Job Open Ecosystem.

The concrete problem it solves is the operational gap left by single-node schedulers. A cron entry or a Quartz-style trigger running on one machine has no answer for what happens when the node dies, when a job needs more throughput than one executor can provide, or when the same job must be split across several workers. ElasticJob shards jobs across a cluster, keeps them highly available, and rebalances automatically when servers are added. Developers no longer have to code around scale out, and operators no longer have to manage high availability by hand; adding servers is enough for the cluster to pick up the work.

Key capabilities

• Elastic Schedule: job sharding and high availability across a distributed system, so throughput scales with the allocation of resources.
• Resource Assign: executes a job at the suitable time on assigned resources, aggregates the same job onto the same executor, and appends resources to newly assigned jobs dynamically.
• Job Governance: failover, misfired job handling, and self-diagnosis with recovery when the distributed environment is unstable.
• Unified job API: code once and deploy at will, with an open architecture for extension and support for Spring IOC.
• Rich job type library: dataflow, script, HTTP, file and big data job types.
• Admin Console, delivered in the separate apache/shardingsphere-elasticjob-ui repository: job administration, job event trace query and registry center management.
• Maven Central artifact org.apache.shardingsphere.elasticjob:elasticjob.

Who users are and how they work

• Java teams running scheduled work across a cluster of executors, where ElasticJob replaces local cron or a single Quartz instance.
• Organizations that already operate ZooKeeper as a coordination service, since ZooKeeper doubles as the registry center.
• Spring-based applications that need scheduled jobs wired through the existing IOC container.
• Operations teams that scale batch throughput by adding servers rather than by rewriting scheduling logic, and that use the Admin Console for job administration and event trace queries.

Getting started

Add the Maven Central artifact org.apache.shardingsphere.elasticjob:elasticjob to a Java project. Building and running require Java 8 or above, Maven 3.5.0 or above, and ZooKeeper 3.6.0 or above as the registry center.

How it compares

The registry lists no paid products that this project replaces. Among similar tools named in the facts, the topics place it next to cron and Quartz, which are single-node scheduling approaches; ElasticJob keeps the familiar scheduled-job model but moves sharding, failover and high availability into a decentralized cluster.

When to use it — and when not to

Self-hosting means operating a ZooKeeper 3.6.0 or above ensemble as the registry center, and building against Java 8 or above with Maven 3.5.0 or above. Teams outside the Java ecosystem, or those unwilling to run ZooKeeper, should look elsewhere. The README is light on step-by-step installation, Job Dependency with DAG based job dependency and DAG based job item dependency is still marked TODO, and the repository carries 116 open issues, so anyone expecting a finished dependency graph should wait.