Parent

Epic: #426

Summary

Deduplicate repetitive log output by normalizing timestamps, UUIDs, hex addresses, and long numbers, then counting identical patterns.

Expected Savings

70-85% on verbose logs.

Behavior

# Before (cargo run output, 500 lines):
2025-01-15T10:00:01Z INFO agent: processing request id=abc-123
2025-01-15T10:00:02Z INFO agent: processing request id=def-456
2025-01-15T10:00:03Z INFO agent: processing request id=ghi-789
... (repeated 200 times)
2025-01-15T10:00:04Z ERROR agent: connection refused addr=127.0.0.1:5432
2025-01-15T10:00:05Z ERROR agent: connection refused addr=127.0.0.1:5432

# After:
[×200] INFO agent: processing request id=<UUID>
[×2] ERROR agent: connection refused addr=<ADDR>

Normalization Rules

• Strip ISO timestamps: \d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}.*?\s
• Replace UUIDs: [0-9a-f]{8}-[0-9a-f]{4}-... → <UUID>
• Replace hex: 0x[0-9a-fA-F]+ → <HEX>
• Replace long numbers: \b\d{4,}\b → <NUM>
• Replace IP:port: \d+\.\d+\.\d+\.\d+:\d+ → <ADDR>

Implementation

• LogDeduplicationFilter — generic, applied when no command-specific filter matches and output has many lines (>50) with high repetition ratio
• HashMap of normalized patterns → (count, first_raw_line)
• Emit in order of first occurrence

Acceptance Criteria

• Deduplication with counts
• Normalization preserves error context
• Only activates on repetitive output (>30% duplication)
• Unit tests