fix(vllm): accept --disable-log-requests as no-op for vLLM 0.17+ compat

[lilyz-ai]

Summary

• --disable-log-requests was removed from vLLM's arg parser in v0.17+
• model-engine hardcodes this flag when launching vllm_server.py, causing startup failures with unrecognized arguments: --disable-log-requests
• Fix: check if the flag already exists before adding it, so this is a no-op when vLLM already handles it and a backward-compat shim when it doesn't

Test plan

• Deployed qwen3-5-27b and qwen3-5-0-8b endpoints with vLLM 0.17.0-rc1 image — both confirmed 3/3 Running

🤖 Generated with Claude Code

Greptile Summary

This PR adds a backward-compatibility shim in vllm_server.py so that the --disable-log-requests CLI flag (hardcoded by model-engine when sensitive_log_mode is enabled) is silently accepted rather than causing an unrecognized arguments startup failure in vLLM 0.17+, where the flag was removed from the arg parser.

Key points:

• The fix correctly uses parser._actions to check whether --disable-log-requests is already registered before adding a no-op handler, so pre-0.17 behavior is fully preserved.
• The main concern is that --disable-log-requests is not merely a legacy artifact — it is still emitted by the current llm_model_endpoint_use_cases.py whenever hmi_config.sensitive_log_mode is True. With this shim in place, vLLM 0.17+ will silently ignore the flag, meaning request logging will not be suppressed for sensitive_log_mode endpoints. If this mode carries privacy or compliance expectations, this is a behavioral regression that should be explicitly acknowledged and tracked.
• The comment describing the shim attributes the flag to "older model-engine versions", which is inaccurate and may mislead future maintainers.

Confidence Score: 3/5

• Safe to merge for startup reliability, but carries an unaddressed behavioral regression for sensitive_log_mode endpoints on vLLM 0.17+.
• The mechanical fix (preventing startup crashes) is correct and tested. However, the PR silently converts an active feature (sensitive_log_mode → --disable-log-requests) into a no-op for vLLM 0.17+ without documenting an alternative or confirming vLLM 0.17+ still suppresses request logs through another mechanism. This warrants attention before merging in compliance-sensitive environments.
• model-engine/model_engine_server/inference/vllm/vllm_server.py — the shim comment is misleading, and the interaction with sensitive_log_mode in llm_model_endpoint_use_cases.py should be explicitly addressed.

Important Files Changed

Filename	Overview
model-engine/model_engine_server/inference/vllm/vllm_server.py	Adds a backward-compat no-op shim for --disable-log-requests removed in vLLM 0.17+. The approach is technically sound but silently disables the behavior of sensitive_log_mode (which relies on this flag to suppress request logs) without a documented replacement path.

Sequence Diagram

sequenceDiagram
    participant ME as model-engine<br/>(llm_model_endpoint_use_cases.py)
    participant SH as Shell
    participant VS as vllm_server.py
    participant PA as FlexibleArgumentParser

    ME->>SH: python -m vllm_server ... --disable-log-requests
    note over ME: sensitive_log_mode=True<br/>→ disable_log_requests=True<br/>→ appends --disable-log-requests to cmd

    SH->>VS: launch with --disable-log-requests in sys.argv

    VS->>PA: make_arg_parser(parser)
    note over PA: vLLM < 0.17: --disable-log-requests registered<br/>vLLM ≥ 0.17: flag NOT registered

    VS->>PA: check parser._actions for --disable-log-requests

    alt vLLM < 0.17 (flag already exists)
        PA-->>VS: flag found → skip add_argument
        PA->>PA: parse_args() — flag works as intended
    else vLLM ≥ 0.17 (flag missing)
        VS->>PA: add_argument("--disable-log-requests", no-op)
        PA->>PA: parse_args() — flag accepted but ignored
        note over PA: ⚠️ request logs NOT suppressed
    end

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: model-engine/model_engine_server/inference/vllm/vllm_server.py
Line: 26-36

Comment:
**Silent no-op may break `sensitive_log_mode` compliance guarantees**

The comment says "older model-engine versions pass `--disable-log-requests`", but the *current* model-engine code (`llm_model_endpoint_use_cases.py:996-997`) still hardcodes this flag whenever `hmi_config.sensitive_log_mode` is `True`:

```python
if hmi_config.sensitive_log_mode:
    vllm_args.disable_log_requests = True
```

By silently accepting `--disable-log-requests` as a no-op in vLLM 0.17+, request logging will **not** be disabled for endpoints that have `sensitive_log_mode` set. If `sensitive_log_mode` is used for privacy or compliance reasons, this silent behavioral regression could expose request-level logs that were previously suppressed.

It would be worth either:
1. Confirming that vLLM 0.17+ disables request logging by default (or via a different mechanism) and documenting that here, or
2. Identifying the replacement mechanism in vLLM 0.17+ (e.g., a new flag or `VLLM_CONFIGURE_LOGGING` env var) and wiring `sensitive_log_mode` through that path alongside this no-op shim.

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: model-engine/model_engine_server/inference/vllm/vllm_server.py
Line: 26-27

Comment:
**Misleading comment — this is current code, not legacy**

The comment attributes the flag to "older model-engine versions", but `--disable-log-requests` is passed by the **current** model-engine code in `llm_model_endpoint_use_cases.py` (conditionally, when `sensitive_log_mode` is enabled). Calling it "older" could mislead future developers into treating this shim as safe to remove once the old caller is gone, when in fact the caller is still active.

```suggestion
    # Compatibility shim: model-engine passes --disable-log-requests (controlled by
    # sensitive_log_mode) which was removed from vLLM's arg parser in v0.17+.
    # Accept it as a no-op to avoid "unrecognized arguments" startup failures.
    # TODO: Investigate whether vLLM 0.17+ exposes an alternative mechanism for
    # suppressing request logs and update llm_model_endpoint_use_cases.py accordingly.
```

**Rule Used:** Add comments to explain complex or non-obvious log... ([source](https://app.greptile.com/review/custom-context?memory=928586f9-9432-435e-a385-026fa49318a2))

**Learnt From**
[scaleapi/scaleapi#126958](https://github.com/scaleapi/scaleapi/pull/126958)

How can I resolve this? If you propose a fix, please make it concise.

_{Last reviewed commit: "fix(vllm): add --dis..."}

Greptile also left 1 inline comment on this PR.

Context used:

• Rule used - Add comments to explain complex or non-obvious log... (source)

Learnt From
scaleapi/scaleapi#126958

[greptile-apps]

Silent no-op may break sensitive_log_mode compliance guarantees

The comment says "older model-engine versions pass --disable-log-requests", but the current model-engine code (llm_model_endpoint_use_cases.py:996-997) still hardcodes this flag whenever hmi_config.sensitive_log_mode is True:

if hmi_config.sensitive_log_mode:
    vllm_args.disable_log_requests = True

By silently accepting --disable-log-requests as a no-op in vLLM 0.17+, request logging will not be disabled for endpoints that have sensitive_log_mode set. If sensitive_log_mode is used for privacy or compliance reasons, this silent behavioral regression could expose request-level logs that were previously suppressed.

It would be worth either:

1. Confirming that vLLM 0.17+ disables request logging by default (or via a different mechanism) and documenting that here, or
2. Identifying the replacement mechanism in vLLM 0.17+ (e.g., a new flag or VLLM_CONFIGURE_LOGGING env var) and wiring sensitive_log_mode through that path alongside this no-op shim.

Prompt To Fix With AI

This is a comment left during a code review.
Path: model-engine/model_engine_server/inference/vllm/vllm_server.py
Line: 26-36

Comment:
**Silent no-op may break `sensitive_log_mode` compliance guarantees**

The comment says "older model-engine versions pass `--disable-log-requests`", but the *current* model-engine code (`llm_model_endpoint_use_cases.py:996-997`) still hardcodes this flag whenever `hmi_config.sensitive_log_mode` is `True`:

```python
if hmi_config.sensitive_log_mode:
    vllm_args.disable_log_requests = True
```

By silently accepting `--disable-log-requests` as a no-op in vLLM 0.17+, request logging will **not** be disabled for endpoints that have `sensitive_log_mode` set. If `sensitive_log_mode` is used for privacy or compliance reasons, this silent behavioral regression could expose request-level logs that were previously suppressed.

It would be worth either:
1. Confirming that vLLM 0.17+ disables request logging by default (or via a different mechanism) and documenting that here, or
2. Identifying the replacement mechanism in vLLM 0.17+ (e.g., a new flag or `VLLM_CONFIGURE_LOGGING` env var) and wiring `sensitive_log_mode` through that path alongside this no-op shim.

How can I resolve this? If you propose a fix, please make it concise.