fix: adjust usage of vLLM for deprecation of 'disable_log_requests'

[jdebache]

Overview:

In vllm-project/vllm#29402, disable_log_requests has been removed. This results in a crash when using Dynamo with the ToT from vLLM, and will result in the same crash when using it with the upcoming 0.12 release.

This PR addresses this by swapping the usages of disable_log_requests by the corresponding usage of enable_log_requests.

Summary by CodeRabbit

• Refactor
  • Updated request logging configuration parameters for the vLLM engine to improve configuration clarity and consistency across components.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[github-actions]

👋 Hi hypdeb! Thank you for contributing to ai-dynamo/dynamo.

Just a reminder: The NVIDIA Test Github Validation CI runs an essential subset of the testing framework to quickly catch errors.Your PR reviewers may elect to test the changes comprehensively before approving your changes.

🚀

[coderabbitai]

Walkthrough

This pull request renames the vLLM configuration parameter from disable_log_requests to enable_log_requests and inverts the corresponding logic by flipping True to False across configuration defaults and function calls in both the core components and example modules.

Changes

Cohort / File(s)	Summary
Dynamo Components Configuration components/src/dynamo/vllm/args.py, components/src/dynamo/vllm/main.py	Updated default configuration from "disable_log_requests": True to "enable_log_requests": False and modified the AsyncLLM initialization call to pass the new parameter with inverted semantics.
Multimodal Examples Configuration examples/multimodal/utils/args.py, examples/multimodal/components/worker.py	Renamed the vLLM engine configuration parameter from disable_log_requests to enable_log_requests and updated corresponding function calls with inverted boolean logic.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

• Verify the logic inversion is consistently applied across all four files (True → False when switching from disable to enable semantics)
• Confirm no additional references to the old disable_log_requests parameter exist elsewhere in the codebase
• Validate that the AsyncLLM API correctly accepts enable_log_requests as a parameter

Poem

🐰 A config rename hops through the code,
Disable becomes enable, a logical mode,
True flips to False with semantic grace,
Log requests now dance in their rightful place! ✨

Pre-merge checks

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 75.00% which is insufficient. The required threshold is 80.00%.	You can run @coderabbitai generate docstrings to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the main change: adjusting vLLM usage due to deprecation of 'disable_log_requests', which directly aligns with the changeset's focus on swapping this parameter across multiple files.
Description check	✅ Passed	The description provides an overview and implementation approach but lacks details section and specific file guidance. However, it includes sufficient context about the issue and the fix approach.

---

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

• X
• Mastodon
• Reddit
• LinkedIn

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

[coderabbitai]

Actionable comments posted: 0

🧹 Nitpick comments (1)

components/src/dynamo/vllm/args.py (1)

450-459: Core vLLM defaults now aligned with enable_log_requests

Updating the Dynamo vLLM defaults to set "enable_log_requests": False keeps the prior “turn off request logging by default” behavior while matching the new vLLM configuration field name. The overwrite loop still enforces these defaults over whatever came from AsyncEngineArgs, which is consistent with the previous design.

If you intend to eventually allow operators to control request logging purely via vLLM CLI flags, you might later relax this to only set enable_log_requests when it is None or unset, mirroring how enable_prefix_caching is handled earlier.

📜 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 c9d7d95 and 89c9fb2.

📒 Files selected for processing (4)

• components/src/dynamo/vllm/args.py (1 hunks)
• components/src/dynamo/vllm/main.py (1 hunks)
• examples/multimodal/components/worker.py (1 hunks)
• examples/multimodal/utils/args.py (1 hunks)

⏰ 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). (1)

• GitHub Check: Build and Test - dynamo

🔇 Additional comments (3)

components/src/dynamo/vllm/main.py (1)

231-237: AsyncLLM call correctly updated to use enable_log_requests

Wiring enable_log_requests=engine_args.enable_log_requests into AsyncLLM.from_vllm_config is consistent with the new vLLM API and with your updated defaults (which now set enable_log_requests=False).

Please confirm against the vLLM version you target that:

• AsyncLLM.from_vllm_config(...) accepts an enable_log_requests keyword, and
• AsyncEngineArgs exposes an enable_log_requests attribute.

If either name differs by version, you may need a small compatibility shim here.

examples/multimodal/components/worker.py (1)

141-147: Multimodal worker correctly switched to enable_log_requests

The multimodal worker’s AsyncLLM.from_vllm_config call now forwards enable_log_requests=self.engine_args.enable_log_requests, matching both the core worker path and your updated defaults in examples.multimodal.utils.args.overwrite_args.

Please verify against your target vLLM version that this parameter name and the corresponding engine_args attribute are present and that there are no remaining call sites still using the deprecated disable_log_requests keyword.

examples/multimodal/utils/args.py (1)

148-171: Migration to enable_log_requests in defaults is semantically correct

Switching the Dynamo multimodal defaults from disable_log_requests=True to enable_log_requests=False preserves the "no per‑request logging by default" behavior while aligning with the new vLLM API. The hasattr + setattr pattern maintains compatibility with AsyncEngineArgs and fails fast if the expected field is missing.

As a follow-up, confirm there are no lingering references to disable_log_requests elsewhere in the codebase to ensure a complete migration.

[rmccorm4]

I think we can remove the setting with the new inverted behavior as described here: #4659 (comment)

But approving to avoid blocking a fix for incompatibility with ToT vLLM

[rmccorm4]

/ok to test 7041619