ARM64-Compatible Images for Sunbird Services (AWS Graviton t4g)

[AbhilashKD]

Hello Team,

We are currently running the following Sunbird/Knowlg platform services on an AWS EC2 t3a.large (x86_64 → 64-bit AMD/Intel) instance.
One of our users requires testing on ARM-based (aarch64 → 64-bit ARM) architecture, specifically AWS Graviton (t4g family).

To proceed with compatibility validation, we need ARM64-supported images for the following services.

🔧 Current x86_64 Images in Use

assessment
sunbirded.azurecr.io/assessment-service:release-8.0.0_ccfcbb4_33

knowledge-platform (Knowlg)
sunbirded.azurecr.io/knowledge-platform-jobs:release-5.7.0_RC3_18c0d23_90

data-pipeline
sunbirded.azurecr.io/data-pipeline:release-7.0.0_RC7_7425f7c_15

content-service
sunbirded.azurecr.io/content-service:release-5.7.0_RC11_a278cff_29

learning-service
shouryasolanki/learning-service:latest

neo4j
sunbirded.azurecr.io/neo4j:3.3.0

print-service
sunbirded.azurecr.io/print-service:release-3.0.0_RC1

search-service
sunbirded.azurecr.io/search-service:release-5.7.0_RC8_7dbe7f8_105

taxonomy-service
sunbirded.azurecr.io/taxonomy-service:release-5.7.0_RC7_a9fa6d9_13

📝 Requirement

We need ARM64-compatible Docker images for all the services listed above so we can deploy and verify functionality on:

AWS EC2 t4g series (Graviton – 64-bit ARM / aarch64)
This is needed to:

Validate ARM compatibility
Perform performance and cost comparison
Support users running ARM-based infrastructure
⏳ Request

Please provide ARM64 images or multi-arch manifests for the above services as soon as possible, so that we can begin our ARM-based testing.

If specific branches or pipelines already support ARM builds, please share those details as well.
Thank you! 🙏
Looking forward to the updated images.

[AbhilashKD]

Any Updates on this ? When can we get this ?

[aimansharief]

Hey @AbhilashKD ,

At the moment, we don’t have pre-built ARM64 images available for these services — the images currently published in our registries support only the x86_64 architecture.

However, all the relevant Dockerfiles and source repositories are public, so you can use them to build ARM64 or multi-arch images and push them to your own Docker/GitHub registries.

Additionally, in our latest releases (8.0.0 or 8.1.0), we have introduced GitHub Actions workflows that build and push images automatically. You can refer to those workflows as a reference for setting up your own ARM64 or multi-arch image builds.

Here are the primary repositories you can refer to:

assessment-service:
https://github.com/Sunbird-inQuiry/inquiry-api-service/tree/release-8.0.0/build

knowledge-platform-jobs
https://github.com/Sunbird-Knowlg/knowledge-platform-jobs/tree/release-5.7.0/jobs-distribution

data-pipeline
https://github.com/Sunbird-inQuiry/data-pipeline/tree/release-7.0.0_RC7

content-service , search-service, taxonomy-service :
https://github.com/Sunbird-Knowlg/knowledge-platform/tree/release-5.7.0_RC7/build

learning-service
https://github.com/Sunbird-Knowlg/sunbird-learning-platform/tree/release-5.7.0/pipelines/build

print-service
https://github.com/project-sunbird/print-service/tree/release-3.0.0

For Neo4j specifically, you can use the official Neo4j base image, build the project from the following repository, and add the generated JARs into the required plugin path inside the image:

Repo: https://github.com/Sunbird-Knowlg/knowledge-platform-db-extensions
This will allow you to produce an ARM64-compatible Neo4j image with the required extensions included

Please let us know if you need any help.

cc: @pallakartheekreddy