Seata Benchmark 1.0 Development Task

[l81893521]

Check Ahead

• [x] I have searched the issues of this repository and believe that this is not a duplicate.

• [ ] I am willing to try to implement this feature myself.

Why you need it?

As a distributed transaction solution, Seata's performance across different transaction modes is a key concern for users. Currently, there lacks a simple, focused tool to stress test various transaction modes (AT, TCC, Saga, etc.) and display critical table data changes in real-time.

How it could be?

Develop a command-line version Benchmark tool capable of stress testing Seata's transaction modes and displaying key information changes from the global transaction table (global_table) and branch transaction table (branch_table) in real-time during testing, enabling visual observation of transaction state transitions and performance metrics

Other related information

No response

[WangzJi]

Create a standalone benchmark module with a CLI tool to stress test Seata transaction modes (AT, TCC, Saga, XA) and display real-time metrics from global_table and branch_table.

Implementation Steps

1. Create Benchmark Module

• Add benchmark module to root pom.xml
• Create benchmark/pom.xml with dependencies: server, core, metrics, JCommander
• Set up package structure

2. CLI Interface (JCommander)

• Command-line options:
  • --mode: Transaction mode (AT/TCC/SAGA/XA/ALL)
  • --storage: Storage mode (db/file/redis/raft)
  • --threads: Number of concurrent threads
  • --duration: Benchmark duration (seconds)
  • --tps-target: Target transactions per second
  • --config: Path to Seata configuration

3. Transaction Simulator

• Mock GlobalSession and BranchSession creation
• Use TransactionStoreManager to write sessions
• Simulate transaction state transitions (Begin → Committing → Committed)
• Support different transaction modes with appropriate branch types

4. Metrics Collection

• Track: TPS, Success Rate, Avg Latency, P95/P99 Latency
• Query global_table for: total count, status distribution
• Query branch_table for: total count, branch type distribution
• Monitor storage layer performance

5. Real-time Console Display

• Live updating dashboard using ANSI escape codes
• Show:
  • Current TPS and target TPS
  • Transaction counts by status (Begin/Committing/Committed/Rollback)
  • Branch counts by type (AT/TCC/SAGA/XA)
  • Table statistics refresh every second
  • Performance metrics (latency percentiles)

6. Storage Adapter

• Abstract interface to work with all storage modes
• Use existing TransactionStoreManager implementations
• Handle configuration for each storage type

Key Technical Details

• Leverage existing classes: GlobalSession, BranchSession, TransactionStoreManager
• Use ServerTableColumnsName constants for table queries
• Support all BranchType enums: AT, TCC, SAGA, XA, SAGA_ANNOTATION
• Thread pool for concurrent transaction generation
• Graceful shutdown to ensure data consistency

This is a plan generated by AI. What do you think about it? If this issue is not urgent, I'd like to give it a try.

[l81893521]

Create a standalone benchmark module with a CLI tool to stress test Seata transaction modes (AT, TCC, Saga, XA) and display real-time metrics from global_table and branch_table. Implementation Steps

1. Create Benchmark Module

• Add benchmark module to root pom.xml
• Create benchmark/pom.xml with dependencies: server, core, metrics, JCommander
• Set up package structure

2. CLI Interface (JCommander)

• Command-line options:

  • --mode: Transaction mode (AT/TCC/SAGA/XA/ALL)
  • --storage: Storage mode (db/file/redis/raft)
  • --threads: Number of concurrent threads
  • --duration: Benchmark duration (seconds)
  • --tps-target: Target transactions per second
  • --config: Path to Seata configuration

3. Transaction Simulator

• Mock GlobalSession and BranchSession creation
• Use TransactionStoreManager to write sessions
• Simulate transaction state transitions (Begin → Committing → Committed)
• Support different transaction modes with appropriate branch types

4. Metrics Collection

• Track: TPS, Success Rate, Avg Latency, P95/P99 Latency
• Query global_table for: total count, status distribution
• Query branch_table for: total count, branch type distribution
• Monitor storage layer performance

5. Real-time Console Display

• Live updating dashboard using ANSI escape codes

• Show:

  • Current TPS and target TPS
  • Transaction counts by status (Begin/Committing/Committed/Rollback)
  • Branch counts by type (AT/TCC/SAGA/XA)
  • Table statistics refresh every second
  • Performance metrics (latency percentiles)

6. Storage Adapter

• Abstract interface to work with all storage modes
• Use existing TransactionStoreManager implementations
• Handle configuration for each storage type

Key Technical Details

• Leverage existing classes: GlobalSession, BranchSession, TransactionStoreManager
• Use ServerTableColumnsName constants for table queries
• Support all BranchType enums: AT, TCC, SAGA, XA, SAGA_ANNOTATION
• Thread pool for concurrent transaction generation
• Graceful shutdown to ensure data consistency

This is a plan generated by AI. What do you think about it? If this issue is not urgent, I'd like to give it a try.

Thank you very much for your response. You are welcome to give it a try. Could you leave your contact information?