refactor: CPU config to millicores, fix replica health, reorder tabs

- Rename cpuShares to cpuRequest (millicores), cpuLimit from cores to
  millicores. ResolvedContainerConfig translates to Docker-native units
  via dockerCpuShares() and dockerCpuQuota() helpers. Future K8s
  orchestrator can pass millicores through directly.
- Fix waitForAnyHealthy to wait for ALL replicas instead of returning
  on first healthy one. Prevents false DEGRADED status with 2+ replicas.
- Default app detail to Configuration tab (was Overview)
- Reorder config sub-tabs: Monitoring, Resources, Variables, Traces &
  Taps, Route Recording

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

cameleer3-server-app/src/main/java/com/cameleer3/server/app/runtime/DeploymentExecutor.java

@@ -37,8 +37,8 @@ public class DeploymentExecutor {
     @Value("${cameleer.runtime.container-memory-limit:512m}")
     private String globalMemoryLimit;
 
-    @Value("${cameleer.runtime.container-cpu-shares:512}")
-    private int globalCpuShares;
+    @Value("${cameleer.runtime.container-cpu-request:500}")
+    private int globalCpuRequest;
 
     @Value("${cameleer.runtime.health-check-timeout:60}")
     private int healthCheckTimeout;
@@ -86,7 +86,7 @@ public class DeploymentExecutor {
 
             var globalDefaults = new ConfigMerger.GlobalRuntimeDefaults(
                     parseMemoryLimitMb(globalMemoryLimit),
-                    globalCpuShares,
+                    globalCpuRequest,
                     globalRoutingMode,
                     globalRoutingDomain,
                     globalServerUrl.isBlank() ? "http://cameleer3-server:8081" : globalServerUrl
@@ -126,7 +126,6 @@ public class DeploymentExecutor {
 
             for (int i = 0; i < config.replicas(); i++) {
                 String containerName = env.slug() + "-" + app.slug() + "-" + i;
-                Long cpuQuota = config.cpuLimit() != null ? (long) (config.cpuLimit() * 100_000) : null;
 
                 String volumeName = jarDockerVolume != null && !jarDockerVolume.isBlank() ? jarDockerVolume : null;
                 ContainerRequest request = new ContainerRequest(
@@ -136,7 +135,7 @@ public class DeploymentExecutor {
                         envNet != null ? List.of(envNet) : List.of(),
                         baseEnvVars, labels,
                         config.memoryLimitBytes(), config.memoryReserveBytes(),
-                        config.cpuShares(), cpuQuota,
+                        config.dockerCpuShares(), config.dockerCpuQuota(),
                         config.exposedPorts(), agentHealthPort,
                         "on-failure", 3
                 );
@@ -273,19 +272,21 @@ public class DeploymentExecutor {
 
     private int waitForAnyHealthy(List<String> containerIds, int timeoutSeconds) {
         long deadline = System.currentTimeMillis() + (timeoutSeconds * 1000L);
+        int lastHealthy = 0;
         while (System.currentTimeMillis() < deadline) {
             int healthy = 0;
             for (String cid : containerIds) {
                 ContainerStatus status = orchestrator.getContainerStatus(cid);
                 if ("healthy".equals(status.state())) healthy++;
             }
-            if (healthy > 0) return healthy;
+            lastHealthy = healthy;
+            if (healthy == containerIds.size()) return healthy;
             try { Thread.sleep(2000); } catch (InterruptedException e) {
                 Thread.currentThread().interrupt();
-                return 0;
+                return lastHealthy;
             }
         }
-        return 0;
+        return lastHealthy;
     }
 
     private List<Map<String, Object>> updateReplicaHealth(List<Map<String, Object>> replicas,