Apache Heron
Apache Heron is a real-time engine for processing data streams across distributed systems, designed to tolerate faults.
What Apache Heron is
Apache Heron is a real-time engine for processing data streams across distributed systems, designed to tolerate faults. Twitter developed it and donated it to the Apache Incubator in 2018. Heron was made compatible with the Apache Storm ecosystem and topology API. Its process-based architecture is written in Java, C++, and Python, with Java and Python Streamlets APIs, a Python API similar to Storm's, and YAML-based ECO topology assembly.
The engine supports backpressure, task isolation for troubleshooting, and configuration for latency-sensitive or throughput-sensitive workloads. It provides stateful, effectively-once processing, with adaptors for Apache Hadoop, Apache BookKeeper, and the local file system. Dhalion is described as helping topologies tune, heal, and stabilize in response to unexpected behavior. Deployment scheduler choices include Kubernetes, DC/OS, Nomad, SLURM, and standalone. Heron was open sourced under Apache License 2.0 and is free for self-hosted Linux use. The project retired on 2023-01-18, and its GitHub repository was archived as read-only on March 3, 2023.
Good
- Supports backpressure and task isolation
- Offers Java and Python APIs
- Includes adaptors for Hadoop and BookKeeper
- Open source under Apache License 2.0
Watch out
- Project retired on 2023-01-18
- GitHub repository is archived read-only
Who it suits
Heron may interest teams maintaining existing stream-processing systems that use its APIs or topology model. Its retired status and read-only repository are important considerations for anyone evaluating it for new work.
Verdict
Heron provides stream-processing capabilities and deployment options, but the project is retired and its repository is read-only. That status is central to deciding whether it fits a current system.
The facts
Each one read from Apache Heron's own pages, with the date it was checked.
What it does
- APIs
- The project describes Java and Python Streamlets APIs, a Python API similar to Storm’s, and YAML-based ECO topology stitching.checked 4 Oct 2026
- Architecture
- Heron uses a process-based computing model with topology isolation.checked 9 Oct 2026
- Archive
- The Apache Heron GitHub repository was archived on March 3, 2023 and is read-only.checked 9 Oct 2026
- Build requirements
- The repository README lists Java 11, Python 3.6, and Bazel 6.0.0 as requirements.checked 9 Oct 2026
- Community support
- The README lists user and development mailing lists and a Heron Slack workspace.checked 9 Oct 2026
- Deployment integrations
- Twitter’s project announcement describes integration with Apache Mesos, Apache Aurora, Apache REEF, and Slurm.checked 9 Oct 2026
- Features
- Heron supports backpressure, task isolation for troubleshooting, and configuring topologies for latency-sensitive or throughput-sensitive workloads.checked 4 Oct 2026
- Intended use
- The repository describes Heron as a real-time stream processing engine developed at Twitter.checked 9 Oct 2026
- Languages and architecture
- Heron uses a process-based architecture and is written in Java, C++ and Python.checked 4 Oct 2026
- License
- The project is licensed under the Apache License, Version 2.0.checked 9 Oct 2026
- Origin
- Twitter developed Heron as a distributed streaming engine and donated it to the Apache Incubator in 2018.checked 4 Oct 2026
- Project status
- The Apache Incubator page says the Heron podling retired on 2023-01-18.checked 9 Oct 2026
- Purpose
- Heron is a real-time, distributed, fault-tolerant stream processing engine.checked 9 Oct 2026
- Schedulers
- The project lists Kubernetes, DC/OS, Nomad, SLURM and standalone as scheduler deployment options.checked 4 Oct 2026
- Security
- The GitHub repository page links to a security policy.checked 9 Oct 2026
- Self-management
- Dhalion is described as enabling Heron to self-tune, self-heal and self-stabilize when topologies encounter unexpected behavior.checked 4 Oct 2026
- State and processing
- The project describes stateful and effectively-once processing, with adaptors for Apache Hadoop, Apache BookKeeper and the local file system.checked 4 Oct 2026
- Stateful processing
- Twitter’s donation announcement describes stateful, exactly-once processing with adaptors for Apache Hadoop, Apache BookKeeper, and the local file system.checked 9 Oct 2026
- Status
- Apache Heron retired on 2023-01-18.checked 4 Oct 2026
- Storm compatibility
- Heron was designed to retain compatibility with Apache Storm’s data model and topology API.checked 9 Oct 2026
- Typical workloads
- Twitter listed ETL, model enhancement, anomaly and fraud detection, IoT, embedded systems, VR/AR, ad bidding, financial, security and social media use cases.checked 4 Oct 2026
- What it does
- Heron is a real-time, distributed, fault-tolerant stream processing engine.checked 4 Oct 2026
Alternatives to Apache Heron
- 1
Apache Beam
Free planOpen sourceWebAPILinuxmacOS7.7Free plan - 2
Apache Druid
Free planWebAPILinuxmacOS7.2Free plan - 3
Confluent Cloud
Free planWebAPILinux7.1From $385 a month - 4
Timeplus
Free planWebAPILinuxmacOS7.1From $199 a month - 5
Materialize
Free planWebAPILinuxSelf-hosted7.0From $1.50 a year - 6
RisingWave
WebAPILinuxmacOS6.7$0.23 usage-based