Skip to content

[Tracking]: Beam 3.0.0 - Milestone 1 Key Features #36173

Closed as not planned
Closed as not planned

Description

@Abacn

What needs to happen?

This umbrella issue tracks key features as part of Beam 3.0.0 milestones

ref: Beam Summit 2025 keynote - https://beamsummit.org/slides/2025/the-future-of-beam.pdf

3 Pillars of Beam 3.0

Access for all

Task sub task timeline (proposed) Notes
Expansive IO Ecosystem
Turnkey transforms
Beam YAML (documentation) First stable release 2.52.0
Structured Data (and transforms)
Schema support improvement (#34672)
CDC / Lakehouse

"Just works" ops

Task sub task timeline (proposed) Notes
Managed Transforms
Introduced Managed Transform API (#30830) and Iceberg Managed IO 2.56.0 Dataflow managed I/O support since Beam 2.58.0
Kafka Managed IO (#31172) 2.57.0
BigQuery Managed IO (#31486) 2.61.0
Jdbc managed IO (#36034) 2.68.0 (postgres)
Fully Featured Local Experience
Prism Runner (#29650) as default direct runner (#34549) 2.50.0 (Go); 2.68.0 (Python)
Whole app tracing
Opentelemetry support (#33176) tbd
Lineage support stable as of Beam 2.66.0 (#34342) OpenLineage support as stretching goal Open (#33981)

First class ML

Task sub task timeline (proposed) Notes
Model Serving
LLM remote inference: Gemini (#35036) and OpenAI (#36081)
Java Remote RunInference (#36253)
RemoteModelHandler API (#34379) and its implementations 2.65.0
Feature Generation
Enrichment support for: CloudSQL (#30773), Milvus (#36176), and Tecton (#36062)
Sinks for: Milvus (#35944), and Tecton (#36062)
Embedding Generation
Multimodal Embedding support: #34236
Embedding ingestion support for: MySQL (#35393), Postgres (#35215), and AlloyDB (#34156)
Anomaly Detection (#35706) Stable as of Beam 2.65.0
New Capabilities
ML-aware API rate limiter

For the changes on existing features/API change for Beam 3.0.0, see SDK trackers:

Activity

  1. damccorm commented on Sep 23, 2025

    @damccorm
    Contributor

    Issues/PRs that stand out to me for First class ML:

    Model Serving

    • Java Remote RunInference - @jrmccluskey do we have an issue for this?
    • Support for LLM remote inference: Gemini and OpenAI

    Feature Generation

    Embedding Generation

    Anomaly Detection

    @claudevdm @jrmccluskey @shunping @tvalentyn could you please add anything that I am missing?

  2. jrmccluskey commented on Sep 23, 2025

    @jrmccluskey
    Contributor

    Tracking issue for Java Remote Inference filed at #36253

    Would we also want to add the model evaluation work here?

  3. damccorm commented on Sep 23, 2025

    @damccorm
    Contributor

    Yes, that's a good idea

  4. liferoad commented on Sep 25, 2025

    @liferoad
    Contributor

    ML-aware API rate limiter on the SDK side could be another work. cc @tarun-google

  5. vampire-hunter-lincoln commented on Dec 15, 2025

    @vampire-hunter-lincoln

    Hi Beam Team,

    Do we have any estimate timeline for Opentelemetry support? I do see a few related PR being merged recently. But I am not sure when Opentelemetry support will be ready.

  6. github-actions commented on May 15, 2026

    @github-actions
    Contributor

    This issue has been marked as stale due to 150 days of inactivity. It will be closed in 30 days if no further activity occurs. If you think that’s incorrect or this issue still needs to be addressed, please simply write any comment. If closed, you can reopen the issue at any time. Thank you for your contributions.

  7. github-actions commented on Jun 15, 2026

    @github-actions
    Contributor

    This issue has been closed due to lack of activity. If you think that is incorrect, you can reopen the issue at any time.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions