fix: misnamed tensorrtllm-runtime image and incorrect tag #4289

saturley-hall · 2025-11-13T01:11:22Z

Overview:

Two incorrect docker image references are corrected here in keeping with standards:

trtllm-runtime is not published on NGC, tensorrtllm-runtime is
The tag dep-554.0 is replaced with my-tag

Details:

The following commands were run (on OSX) to effectuate this change:

rg -l trtllm-runtime --null | xargs -0 sed -i '' 's/trtllm-runtime/tensorrtllm-runtime/g'
rg -l "dep-554.0" --null  | xargs -0 sed -i '' 's/dep-554.0/my-tag/g'

Summary by CodeRabbit

Documentation
- Updated SLA-driven profiling documentation, deployment examples, and prerequisites with current container image naming conventions.
Chores
- Updated container image references across profiling benchmarks, deployment configurations, and example templates.
- Refreshed profiling configuration samples and deployment manifests with revised image specifications.

Signed-off-by: Harrison King Saturley-Hall <[email protected]>

coderabbitai · 2025-11-13T01:16:36Z

Walkthrough

Container image references are updated across deployment configurations and documentation. The image name trtllm-runtime is changed to tensorrtllm-runtime, and image tags are updated from dep-554.0 to my-tag in profiling benchmarks. These changes are applied consistently to YAML manifests and documentation files.

Changes

Cohort / File(s)	Change Summary
Profiling Benchmark Configurations `benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml`, `benchmarks/profiler/deploy/profile_sla_dgdr.yaml`	Image references updated: `profilerImage` and `workersImage` changed from `dep-554.0` to `my-tag` tag; `trtllm-runtime` image name changed to `tensorrtllm-runtime` in profile_sla_aic_dgdr.yaml
Operator Sample Configuration `deploy/cloud/operator/config/samples/nvidia.com_v1alpha1_dynamographdeploymentrequest.yaml`	Profiler container image updated from `trtllm-runtime:0.6.1` to `tensorrtllm-runtime:0.6.1`
Documentation `docs/benchmarks/sla_driven_profiling.md`, `examples/backends/trtllm/deploy/README.md`, `recipes/README.md`	Image references updated in examples and documentation from `trtllm-runtime` to `tensorrtllm-runtime` across SLA profiling docs, deployment README, and recipe prerequisites
TensorRT-LLM Deployment Manifests `examples/backends/trtllm/deploy/agg.yaml`, `agg_router.yaml`, `disagg-multinode.yaml`, `disagg.yaml`, `disagg_planner.yaml`, `disagg_router.yaml`	Container image references updated from `trtllm-runtime:my-tag` to `tensorrtllm-runtime:my-tag` across Frontend and worker component mainContainer specifications
Recipe Deployment Configuration `recipes/gpt-oss-120b/trtllm/agg/deploy.yaml`	Image references updated from `trtllm-runtime:my-tag` to `tensorrtllm-runtime:my-tag` in Frontend and TrtllmWorker service configurations

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Verify image name consistency across all YAML manifests (ensures no typos in the new image name)
Confirm tag values match intended deployment targets (verify my-tag and 0.6.1 are correct)
Check that documentation examples align with actual deployment configurations

Poem

🐰 With whiskers twitching and a hop so spry,
I've updated images, reaching toward the sky,
From trtllm to tensorrtllm they now stand tall,
New tags in place—a rename for them all!
Hopping through configs, both yaml and doc,
Consistent and steady, tick-tock, tick-tock! ✨

Pre-merge checks

✅ Passed checks (3 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main changes: correcting a misnamed Docker image (trtllm-runtime → tensorrtllm-runtime) and fixing an incorrect tag (dep-554.0 → my-tag).
Description check	✅ Passed	The description provides an overview of the changes and technical details on how they were applied, but lacks the 'Where should the reviewer start?' and 'Related Issues' sections from the template.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

📜 Recent review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 3fa8448 and 39acf52.

📒 Files selected for processing (13)

benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml (2 hunks)
benchmarks/profiler/deploy/profile_sla_dgdr.yaml (2 hunks)
deploy/cloud/operator/config/samples/nvidia.com_v1alpha1_dynamographdeploymentrequest.yaml (1 hunks)
docs/benchmarks/sla_driven_profiling.md (3 hunks)
examples/backends/trtllm/deploy/README.md (3 hunks)
examples/backends/trtllm/deploy/agg.yaml (2 hunks)
examples/backends/trtllm/deploy/agg_router.yaml (2 hunks)
examples/backends/trtllm/deploy/disagg-multinode.yaml (3 hunks)
examples/backends/trtllm/deploy/disagg.yaml (3 hunks)
examples/backends/trtllm/deploy/disagg_planner.yaml (4 hunks)
examples/backends/trtllm/deploy/disagg_router.yaml (3 hunks)
recipes/README.md (1 hunks)
recipes/gpt-oss-120b/trtllm/agg/deploy.yaml (2 hunks)

🧰 Additional context used

🧠 Learnings (2)

📚 Learning: 2025-09-04T19:03:06.643Z

Learnt from: biswapanda
Repo: ai-dynamo/dynamo PR: 2872
File: examples/multimodal/deploy/agg_qwen.yaml:53-60
Timestamp: 2025-09-04T19:03:06.643Z
Learning: In the dynamo repository, Kubernetes Custom Resources use `gpu: "1"` format for GPU resource limits and requests, not the standard Kubernetes `nvidia.com/gpu: 1` format. This applies to DynamoGraphDeployment resources and other dynamo CRs.

Applied to files:

benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml
benchmarks/profiler/deploy/profile_sla_dgdr.yaml
deploy/cloud/operator/config/samples/nvidia.com_v1alpha1_dynamographdeploymentrequest.yaml

📚 Learning: 2025-07-31T11:26:48.422Z

Learnt from: KrishnanPrash
Repo: ai-dynamo/dynamo PR: 2217
File: components/backends/trtllm/engine_configs/deepseek_r1/wide_ep/wide_ep_prefill.yaml:18-0
Timestamp: 2025-07-31T11:26:48.422Z
Learning: TRTLLM LLM-API expects all caps for backend field names in configuration files. When migrating TRTLLM configurations, backend values like "WideEP" should be changed to "WIDEEP" to comply with the API requirements.

Applied to files:

examples/backends/trtllm/deploy/disagg_planner.yaml
examples/backends/trtllm/deploy/README.md

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (5)

GitHub Check: vllm (amd64)
GitHub Check: operator (amd64)
GitHub Check: sglang (amd64)
GitHub Check: trtllm (amd64)
GitHub Check: Build and Test - dynamo

🔇 Additional comments (13)

examples/backends/trtllm/deploy/disagg.yaml (1)

16-16: Image references updated consistently.

The container image references across all three services have been updated from the non-published trtllm-runtime to tensorrtllm-runtime with the correct tag maintained.

Also applies to: 28-28, 54-54

deploy/cloud/operator/config/samples/nvidia.com_v1alpha1_dynamographdeploymentrequest.yaml (1)

28-28: Profiler image reference updated to published name.

The profilerImage reference correctly uses tensorrtllm-runtime instead of the non-published trtllm-runtime, with the version tag 0.6.1 preserved.

examples/backends/trtllm/deploy/disagg_planner.yaml (1)

16-16: All four service image references updated correctly.

The container images for Frontend, Planner, Decode, and Prefill worker components have been consistently updated to use tensorrtllm-runtime with the correct tag.

Also applies to: 41-41, 92-92, 119-119

examples/backends/trtllm/deploy/disagg_router.yaml (1)

16-16: Router deployment image references updated consistently.

All three services (Frontend, Prefill, Decode workers) have been updated to use the correct tensorrtllm-runtime image with router configuration preserved.

Also applies to: 30-30, 56-56

examples/backends/trtllm/deploy/disagg-multinode.yaml (1)

98-98: Multinode deployment image references updated.

All three service images in the tensor-parallel (TP8) multinode configuration have been correctly updated to tensorrtllm-runtime, with all multinode, volume, and GPU configuration preserved.

Also applies to: 130-130, 168-168

benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml (1)

15-15: Profiling deployment image references and tags updated.

Both profiler and worker images have been correctly updated with the new tag my-tag, and the workers image now uses the correct tensorrtllm-runtime instead of the non-published trtllm-runtime.

Also applies to: 34-34

recipes/gpt-oss-120b/trtllm/agg/deploy.yaml (1)

49-49: GPT-OSS recipe image references updated.

Both Frontend and TrtllmWorker images in the aggregated mode deployment have been correctly updated to tensorrtllm-runtime, with all tensor parallelism (TP4 EP4), model paths, and resource configurations preserved.

Also applies to: 84-84

examples/backends/trtllm/deploy/agg_router.yaml (1)

16-16: Aggregated router deployment image references updated.

Both Frontend and worker images have been correctly updated to tensorrtllm-runtime with router configuration and resource specifications preserved.

Also applies to: 30-30

recipes/README.md (1)

77-77: Image reference update looks good.

The reference is correctly updated to use the published image name tensorrtllm-runtime instead of the non-published trtllm-runtime.

examples/backends/trtllm/deploy/README.md (1)

92-92: Image references correctly updated across documentation examples.

All references to the non-published trtllm-runtime image are consistently replaced with tensorrtllm-runtime in both code examples and documentation text.

Also applies to: 112-112, 144-144

docs/benchmarks/sla_driven_profiling.md (1)

395-395: Documentation examples correctly updated with published image names.

All three instances in YAML configuration examples now reference tensorrtllm-runtime instead of the non-published trtllm-runtime, ensuring documentation users deploy with the correct images.

Also applies to: 412-412, 496-496

benchmarks/profiler/deploy/profile_sla_dgdr.yaml (1)

15-15: Tag replacement applied consistently, but verify tag naming is intentional.

Both profilerImage and workersImage tags are correctly updated from dep-554.0 to my-tag. However, please confirm that my-tag is the intended final tag name and not a placeholder that should reference an actual version number.

Also applies to: 31-31

examples/backends/trtllm/deploy/agg.yaml (1)

16-16: Deployment manifest images correctly updated for both services.

Both the Frontend and TRTLLMWorker container images are updated to reference the published tensorrtllm-runtime image, ensuring this manifest deploys with the correct runtime.

Also applies to: 27-27

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

nv-anants

lgtm!

hhzhang16 · 2025-11-13T23:28:24Z

benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml

  # ProfilingConfig maps directly to the profile_sla.py config format
  profilingConfig:
-    profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:dep-554.0"
+    profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"


Do you mind changing these to nvidia/ai-dynamo as well?

Suggested change

profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"

profilerImage: "nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag"

hhzhang16 · 2025-11-13T23:28:48Z

benchmarks/profiler/deploy/profile_sla_aic_dgdr.yaml

  # Deployment overrides for the auto-created DGD
  deploymentOverrides:
-    workersImage: "nvcr.io/nvidian/dynamo-dev/trtllm-runtime:dep-554.0"
+    workersImage: "nvcr.io/nvidian/dynamo-dev/tensorrtllm-runtime:my-tag"


Suggested change

workersImage: "nvcr.io/nvidian/dynamo-dev/tensorrtllm-runtime:my-tag"

workersImage: "nvcr.io/nvidia/ai-dynamo/tensorrtllm-runtime:my-tag"

hhzhang16 · 2025-11-13T23:29:08Z

benchmarks/profiler/deploy/profile_sla_dgdr.yaml

  # ProfilingConfig maps directly to the profile_sla.py config format
  profilingConfig:
-    profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:dep-554.0"
+    profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"


Suggested change

profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"

profilerImage: "nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag"

hhzhang16 · 2025-11-13T23:29:18Z

benchmarks/profiler/deploy/profile_sla_dgdr.yaml

  # Deployment overrides for the auto-created DGD
  deploymentOverrides:
-    workersImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:dep-554.0"
+    workersImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"


Suggested change

workersImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"

workersImage: "nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag"

fix: misnamed tensorrtllm-runtime image and incorrect tag

39acf52

Signed-off-by: Harrison King Saturley-Hall <[email protected]>

saturley-hall requested review from athreesh and dagil-nvidia November 13, 2025 01:11

saturley-hall requested review from a team as code owners November 13, 2025 01:11

pull-request-size bot added the size/M label Nov 13, 2025

github-actions bot added the fix label Nov 13, 2025

saturley-hall requested a review from nv-anants November 13, 2025 16:49

nv-anants approved these changes Nov 13, 2025

View reviewed changes

saturley-hall enabled auto-merge (squash) November 13, 2025 16:55

dagil-nvidia approved these changes Nov 13, 2025

View reviewed changes

saturley-hall disabled auto-merge November 13, 2025 19:28

hhzhang16 reviewed Nov 13, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

fix: misnamed tensorrtllm-runtime image and incorrect tag #4289

fix: misnamed tensorrtllm-runtime image and incorrect tag #4289

saturley-hall commented Nov 13, 2025 •

edited by coderabbitai bot

Loading

Uh oh!

coderabbitai bot commented Nov 13, 2025

Uh oh!

nv-anants left a comment

Uh oh!

hhzhang16 Nov 13, 2025

Uh oh!

hhzhang16 Nov 13, 2025

Uh oh!

hhzhang16 Nov 13, 2025

Uh oh!

hhzhang16 Nov 13, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

	profilerImage: "nvcr.io/nvidian/dynamo-dev/vllm-runtime:my-tag"
	profilerImage: "nvcr.io/nvidia/ai-dynamo/vllm-runtime:my-tag"

	workersImage: "nvcr.io/nvidian/dynamo-dev/tensorrtllm-runtime:my-tag"
	workersImage: "nvcr.io/nvidia/ai-dynamo/tensorrtllm-runtime:my-tag"

fix: misnamed tensorrtllm-runtime image and incorrect tag #4289

Are you sure you want to change the base?

fix: misnamed tensorrtllm-runtime image and incorrect tag #4289

Conversation

saturley-hall commented Nov 13, 2025 • edited by coderabbitai bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Overview:

Details:

Summary by CodeRabbit

Uh oh!

coderabbitai bot commented Nov 13, 2025

Walkthrough

Changes

Estimated code review effort

Poem

Pre-merge checks

Uh oh!

nv-anants left a comment

Choose a reason for hiding this comment

Uh oh!

hhzhang16 Nov 13, 2025

Choose a reason for hiding this comment

Uh oh!

hhzhang16 Nov 13, 2025

Choose a reason for hiding this comment

Uh oh!

hhzhang16 Nov 13, 2025

Choose a reason for hiding this comment

Uh oh!

hhzhang16 Nov 13, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

saturley-hall commented Nov 13, 2025 •

edited by coderabbitai bot

Loading