[Bug] Fix NPE when recording RPC latency for local proxy requests

[majialoong]

Before Creating the Bug Report

• I found a bug, not just asking a question, which should be created in GitHub Discussions.

• I have searched the GitHub Issues and GitHub Discussions of this repository and believe that this is not a duplicate.

• I have confirmed that this bug belongs to the current repository, not other repositories of RocketMQ.

Runtime platform environment

All

RocketMQ version

5.x

JDK Version

All

Describe the Bug

In Local Proxy mode, LocalRemotingCommand does not initialize processTimer because the request does not pass through NettyDecoder.

This causes a NullPointerException in the ChannelFutureListener of NettyRemotingAbstract.writeResponse when recording the RPC latency.

Steps to Reproduce

1. Start RocketMQ Proxy in Local mode.
2. Send a message through the gRPC endpoint.
3. Inspect NettyRemotingAbstract.writeResponse when the response is written.
4. Observe that request.getProcessTimer() is null.

What Did You Expect to See?

The local request should have an initialized process timer, and the RPC latency metric should be recorded successfully.

What Did You See Instead?

A NullPointerException occurs when calling request.getProcessTimer().elapsed(...).

Additional Context

No response

[RockteMQ-AI]

Issue Evaluation

Category: type/bug | Status: Confirmed

The reported issue has been verified against the current codebase.

Root Cause: processTimer is declared in RemotingCommand (line 100) but never initialized in the constructor. The only initialization point is NettyDecoder (line 42/49), which creates a Stopwatch.createStarted() and sets it on the decoded command. LocalRemotingCommand bypasses the Netty pipeline entirely — it is constructed in-process by the proxy layer, so processTimer remains null.

Crash Sites: NettyRemotingAbstract.writeResponse dereferences request.getProcessTimer().elapsed(...) unconditionally at 3 locations (lines 250, 267, 279), all without null guards.

Impact: Any gRPC request through Local Proxy mode will trigger NPE when the response is written back, affecting RPC latency metrics recording and potentially crashing the response path.

Severity: High — affects all Local Proxy mode users.

Suggested Fix: Initialize processTimer in RemotingCommand constructor, and/or add null guards at the writeResponse call sites.

An automated fix proposal will be generated. Reply /approve to proceed with PR generation.

---

Automated evaluation by github-manager-bot