fix: remove log tail limit in validation script for model loading detection

The validation script was failing to detect successful model loading because
it used --tail=200 to check container logs. Since model containers generate
many health check logs after startup, the "model loaded" messages were
beyond the 200-line limit.

Changes:
- Remove --tail=200 from model-a validation (line 98)
- Remove --tail=200 from model-b validation (line 110)
- Now scans entire log history for "model loaded" message

Fixes #391

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <[email protected]>
Signed-off-by: Yossi Ovadia <[email protected]>

Author: yossiovadia

deploy/openshift/validate-deployment.sh

@@ -95,7 +95,7 @@ fi
 # Test 6: Check model-a loaded successfully
 log "INFO" "Test 6: Checking model-a loaded successfully..."
 if [[ -n "$POD_NAME" ]]; then
-    MODEL_STATUS=$(oc logs -n "$NAMESPACE" "$POD_NAME" -c model-a --tail=200 | grep -i "model loaded" | tail -1)
+    MODEL_STATUS=$(oc logs -n "$NAMESPACE" "$POD_NAME" -c model-a | grep -i "model loaded" | tail -1)
     if [[ -n "$MODEL_STATUS" ]]; then
         DEVICE=$(echo "$MODEL_STATUS" | grep -oE "(cuda|cpu)" || echo "unknown")
         log "PASS" "Model-A loaded successfully on $DEVICE"
@@ -107,7 +107,7 @@ fi
 # Test 7: Check model-b loaded successfully
 log "INFO" "Test 7: Checking model-b loaded successfully..."
 if [[ -n "$POD_NAME" ]]; then
-    MODEL_STATUS=$(oc logs -n "$NAMESPACE" "$POD_NAME" -c model-b --tail=200 | grep -i "model loaded" | tail -1)
+    MODEL_STATUS=$(oc logs -n "$NAMESPACE" "$POD_NAME" -c model-b | grep -i "model loaded" | tail -1)
     if [[ -n "$MODEL_STATUS" ]]; then
         DEVICE=$(echo "$MODEL_STATUS" | grep -oE "(cuda|cpu)" || echo "unknown")
         log "PASS" "Model-B loaded successfully on $DEVICE"