Artemis: feat: Add extensive configuration options for cache, batching, and connection pools

src/main/resources/application.properties

@@ -9,6 +9,56 @@ logging.level.com.llmproxy=DEBUG
 cache.enabled=true
 cache.ttl.seconds=300
 cache.max-items=1000
+cache.eviction-policy=W-TinyLFU
+# W-TinyLFU specific settings
+cache.window-size-percentage=0.01
+cache.sample-size=10
+# LRU/LFU common settings
+cache.initial-capacity=100
+# Weight-based settings
+cache.weight-based=true
+cache.maximum-weight=100000000
+cache.weigher=RESPONSE_SIZE
+# Segment settings for concurrency
+cache.concurrency-level=4
+# Statistics collection
+cache.record-stats=true
+# Per-provider settings
+cache.openai.ttl.seconds=600
+cache.gemini.ttl.seconds=300
+cache.mistral.ttl.seconds=450
+cache.claude.ttl.seconds=500
+# Access-based expiration
+cache.expire-after-access=true
+cache.access-expiration.seconds=1800
+
+# Batch processing configuration (for multi-request grouping/dispatch)
+batch.enabled=true
+batch.max-size=16                # Default max batch size (requests) for all providers
+batch.timeout.ms=50              # Default max wait time (ms) before dispatching non-full batch
+batch.buffer.max-pending=128     # Default buffer size for pending requests per provider
+batch.flush-on-overflow=true     # Flush batch when buffer is full, even before timeout
+
+# Per-provider overrides (if omitted, global defaults apply)
+batch.openai.enabled=true
+batch.openai.max-size=8
+batch.openai.timeout.ms=40
+batch.openai.buffer.max-pending=64
+
+batch.gemini.enabled=true
+batch.gemini.max-size=16
+batch.gemini.timeout.ms=60
+batch.gemini.buffer.max-pending=128
+
+batch.mistral.enabled=true
+batch.mistral.max-size=10
+batch.mistral.timeout.ms=50
+batch.mistral.buffer.max-pending=80
+
+batch.claude.enabled=true
+batch.claude.max-size=12
+batch.claude.timeout.ms=70
+batch.claude.buffer.max-pending=96
 
 # Retry Configuration
 retry.max-attempts=3
@@ -26,3 +76,46 @@ api.openai.key=${OPENAI_API_KEY:}
 api.gemini.key=${GEMINI_API_KEY:}
 api.mistral.key=${MISTRAL_API_KEY:}
 api.claude.key=${CLAUDE_API_KEY:}
+
+# Global connection pool settings
+connection.pool.default.max-total=100
+connection.pool.default.max-per-route=20
+connection.pool.default.connect-timeout-ms=5000
+connection.pool.default.connection-request-timeout-ms=5000
+connection.pool.default.socket-timeout-ms=30000
+connection.pool.default.time-to-live-ms=300000
+connection.pool.default.validate-after-inactivity-ms=10000
+connection.pool.default.evict-idle-connections=true
+connection.pool.default.idle-timeout-ms=60000
+
+# OpenAI connection pool settings
+connection.pool.openai.max-total=50
+connection.pool.openai.max-per-route=20
+connection.pool.openai.connect-timeout-ms=5000
+connection.pool.openai.connection-request-timeout-ms=5000
+connection.pool.openai.socket-timeout-ms=60000
+connection.pool.openai.time-to-live-ms=300000
+
+# Gemini connection pool settings
+connection.pool.gemini.max-total=50
+connection.pool.gemini.max-per-route=20
+connection.pool.gemini.connect-timeout-ms=5000
+connection.pool.gemini.connection-request-timeout-ms=5000
+connection.pool.gemini.socket-timeout-ms=60000
+connection.pool.gemini.time-to-live-ms=300000
+
+# Mistral connection pool settings
+connection.pool.mistral.max-total=50
+connection.pool.mistral.max-per-route=20
+connection.pool.mistral.connect-timeout-ms=5000
+connection.pool.mistral.connection-request-timeout-ms=5000
+connection.pool.mistral.socket-timeout-ms=45000
+connection.pool.mistral.time-to-live-ms=300000
+
+# Claude connection pool settings
+connection.pool.claude.max-total=50
+connection.pool.claude.max-per-route=20
+connection.pool.claude.connect-timeout-ms=5000
+connection.pool.claude.connection-request-timeout-ms=5000
+connection.pool.claude.socket-timeout-ms=90000
+connection.pool.claude.time-to-live-ms=300000