每个RocketMQMessageListener会创建2 * CPU核数的线程

[GavinLi2019]

模块：rocketmq-v5-client-spring-boot

通过注解创建 RocketMQMessageListener 时，通过 registerContainer 方法注册容器，每个容器会创建 ClientImpl
初始化时会创建两个 CPU 核数的线程池对象，
当 12核 CPU ， 30 个 Listener 初始化时，会大约创建 720 个线程

[GavinLi2019]

1个Listener 会创建很多的线程池

[oss-sentinel-ai]

Issue Evaluation

Type: Bug / Performance Issue
Category: Resource Management
Priority: Medium

Assessment

Valid performance concern in rocketmq-v5-client-spring-boot. Each RocketMQMessageListener creates 2 × CPU cores threads via ClientImpl initialization. With 30 listeners on a 12-core CPU, ~720 threads are created.

Impact

• Excessive thread count and memory consumption
• Thread scheduling overhead
• Potential resource exhaustion

✅ Accepted — Thread pool should be shared across listeners or made configurable.

Suggested approaches:

1. Share thread pool across ClientImpl instances
2. Make thread pool size configurable per listener
3. Implement thread pool reuse strategy

---

🤖 Automated evaluation by oss-sentinel-ai

[RockteMQ-AI]

Issue Evaluation

Category: bug / enhancement | Status: Confirmed

Analysis: Each RocketMQMessageListener creates a ClientImpl that initializes a thread pool with 2 * CPU_cores threads. With 30 listeners on a 12-core CPU, this creates ~720 threads, which is excessive and can lead to:

• High memory consumption
• Context switching overhead
• Potential thread exhaustion under load

Suggested fix:

• Share a common ClientManager or thread pool across multiple listeners
• Make the thread pool size configurable per listener or globally
• Consider lazy initialization of thread pools

Workaround: Reduce the number of listeners or use a shared client instance.

---

Automated evaluation by RockteMQ-AI

[RockteMQ-AI]

Issue Evaluation

Category: bug | Status: Confirmed

Root Cause: Each RocketMQMessageListener creates its own ClientImpl instance via registerContainer, which initializes thread pools sized at 2 × CPU cores. With N listeners, total threads = N × 2 × cores.

Impact: On a 12-core machine with 30 listeners → 720 threads. This causes excessive context switching, memory overhead, and potential thread exhaustion.

Severity: Medium-High — affects any application with multiple listeners on multi-core machines.

Suggested fix: Listeners should share a common ClientManager or thread pool rather than each creating independent ClientImpl instances with their own thread pools.

---

Automated evaluation by github-manager

[RockteMQ-AI]

Issue Evaluation

Category: bug | Status: Confirmed

This is a valid resource management issue. Each RocketMQMessageListener creates a ClientImpl instance that initializes thread pools sized at 2 * CPU cores. With N listeners on a machine with C cores, this creates 2 * C * N threads — excessive and wasteful.

Root Cause: The v5 client Spring Boot starter creates a separate ClientImpl per listener instead of sharing a common client instance across listeners.

Impact: On a 12-core machine with 30 listeners: 720 threads just for client initialization. This wastes memory and creates unnecessary context-switching overhead.

Severity: medium

Suggested fix: Share a single ClientImpl instance across all listeners in the same application context, or use a connection pool pattern.

---

Automated evaluation by RockteMQ-AI

[RockteMQ-AI]

Issue Evaluation

Category: bug | Status: Confirmed

Valid resource concern. Each RocketMQMessageListener creates a ClientImpl which initializes thread pools sized at 2× CPU cores. With 30 listeners on a 12-core machine, this creates ~720 threads — excessive and wasteful.

Root Cause: Per-listener ClientImpl instantiation without shared thread pool or resource pooling.
Impact: Excessive thread creation, increased memory usage, and potential thread exhaustion under high listener counts.
Severity: Medium-High

Suggested approach: Share a common ClientImpl (or at least its thread pools) across multiple listeners, or use a bounded thread pool that scales with actual workload rather than CPU cores per listener.

---

Automated evaluation by github-manager

[RockteMQ-AI]

Issue Evaluation\n\nCategory: question | Status: Needs Clarification\n\nThis appears to be a usage question or configuration issue. Please provide:\n- RocketMQ Spring version\n- Configuration details\n- Error logs or stack traces if applicable\n\nFor usage questions, consider checking the documentation or asking in GitHub Discussions.\n\n---\nAutomated evaluation by github-manager

[RockteMQ-AI]

Issue Evaluation

Category: bug | Status: Confirmed

Each RocketMQMessageListener creates a ClientImpl with 2×CPU-core thread pools. With N listeners on a 12-core machine, this creates 24N threads — clearly excessive.

Root Cause: No thread pool sharing between listener instances; each creates independent ClientImpl.
Impact: Resource exhaustion with many listeners. On a 12-core machine with 30 listeners = ~720 threads.
Severity: High

This should be addressed by sharing a common ClientImpl or thread pool across listeners.

---

Automated evaluation by RockteMQ-AI

[RockteMQ-AI]

Issue Evaluation

Category: enhancement | Status: Evaluated

Feasibility: needs-discussion
Scope: RocketMQMessageListener thread pool configuration
Compatibility: May affect existing deployments that rely on current thread allocation

Creating 2 * CPU cores threads per listener can lead to excessive thread creation when multiple listeners are configured. A shared, configurable thread pool would be more efficient.

---

Automated evaluation by github-manager-bot

[RockteMQ-AI]

Issue Evaluation

Category: bug | Status: Confirmed

Creating 2×CPU threads per @RocketMQMessageListener is excessive and can lead to thread exhaustion when multiple listeners are configured. This is a resource management issue that should be addressed.

Impact: Performance and resource consumption
Severity: Medium-High

Consider implementing thread pool sharing across listeners or making thread count configurable.

---

Automated evaluation by RockteMQ-AI

[RockteMQ-AI]

Issue Evaluation

Category: type/bug | Status: Confirmed

The thread multiplication issue is real. Each RocketMQMessageListener creates a ClientImpl that initializes two thread pools sized by CPU cores. With 30 listeners on a 12-core machine, this creates ~720 threads — far more than necessary.

The fix should share a common ClientImpl (or at least its thread pools) across listeners that connect to the same cluster, rather than creating independent instances per listener. This is a resource efficiency issue that affects deployments with many listeners.

---

Automated evaluation by github-manager