fix: add instance-dead-timeout configuration and update HealthMonitor to use it for stale instance eviction

modules/coordinator/src/main/resources/application.yml

@@ -33,6 +33,7 @@ nowchess:
     rebalance-interval: 30s
     rebalance-min-interval: 60s
     heartbeat-ttl: 5s
+    instance-dead-timeout: 60s
     stream-heartbeat-interval: PT0.2S
     cache-eviction-interval: 10m
     game-idle-threshold: 45m

modules/coordinator/src/main/scala/de/nowchess/coordinator/config/CoordinatorConfig.scala

@@ -21,6 +21,9 @@ trait CoordinatorConfig:
   @WithName("heartbeat-ttl")
   def heartbeatTtl: Duration
 
+  @WithName("instance-dead-timeout")
+  def instanceDeadTimeout: Duration
+
   @WithName("stream-heartbeat-interval")
   def streamHeartbeatInterval: Duration
 

modules/coordinator/src/main/scala/de/nowchess/coordinator/service/AutoScaler.scala

@@ -50,8 +50,7 @@ class AutoScaler:
           val avgLoad = instances.map(_.subscriptionCount).sum.toDouble / instances.size
 
           if avgLoad > config.scaleUpThreshold * config.maxGamesPerCore then scaleUp()
-          else if avgLoad < config.scaleDownThreshold * config.maxGamesPerCore && instances.size > config.scaleMinReplicas
+          else if avgLoad < config.scaleDownThreshold * config.maxGamesPerCore then scaleDown()
-          then scaleDown()
 
   def scaleUp(): Unit =
     log.info("Scaling up Argo Rollout")

modules/coordinator/src/main/scala/de/nowchess/coordinator/service/HealthMonitor.scala

@@ -39,7 +39,7 @@ class HealthMonitor:
     redisPrefix = prefix
 
   def checkInstanceHealth: Unit =
-    val evicted = instanceRegistry.evictStaleInstances(config.heartbeatTtl)
+    val evicted = instanceRegistry.evictStaleInstances(config.instanceDeadTimeout)
     if evicted.nonEmpty then log.warnf("Evicted %d stale instances: %s", evicted.size, evicted.mkString(", "))
     val instances = instanceRegistry.getAllInstances
     instances.foreach { inst =>