removed pods_in_node and utilizing max_pods

modules/python/clusterloader2/slo/slo.py

@@ -24,7 +24,7 @@
 }
 # TODO: Remove aks once CL2 update provider name to be azure
 
-def calculate_config(cpu_per_node, node_count, pods_in_node, provider, service_test, cnp_test, ccnp_test):
+def calculate_config(cpu_per_node, node_count, max_pods, provider, service_test, cnp_test, ccnp_test):
     throughput = 100
     nodes_per_namespace = min(node_count, DEFAULT_NODES_PER_NAMESPACE)
 
@@ -33,7 +33,7 @@ def calculate_config(cpu_per_node, node_count, pods_in_node, provider, service_t
         pods_per_node = LOAD_PODS_PER_NODE
 
     if cnp_test or ccnp_test:
-        pods_per_node = pods_in_node
+        pods_per_node = max_pods
     # Different cloud has different reserved values and number of daemonsets
     # Using the same percentage will lead to incorrect nodes number as the number of nodes grow
     # For AWS, see: https://github.com/awslabs/amazon-eks-ami/blob/main/templates/al2/runtime/bootstrap.sh#L290
@@ -49,7 +49,6 @@ def configure_clusterloader2(
     node_count,
     node_per_step,
     max_pods,
-    pods_in_node,
     repeats,
     operation_timeout,
     provider,
@@ -63,7 +62,7 @@ def configure_clusterloader2(
     override_file):
 
     steps = node_count // node_per_step
-    throughput, nodes_per_namespace, pods_per_node, cpu_request = calculate_config(cpu_per_node, node_per_step, pods_in_node, provider, service_test, cnp_test, ccnp_test)
+    throughput, nodes_per_namespace, pods_per_node, cpu_request = calculate_config(cpu_per_node, node_per_step, max_pods, provider, service_test, cnp_test, ccnp_test)
 
     with open(override_file, 'w') as file:
         file.write(f"CL2_LOAD_TEST_THROUGHPUT: {throughput}\n")
@@ -131,7 +130,6 @@ def collect_clusterloader2(
     cpu_per_node,
     node_count,
     max_pods,
-    pods_in_node,
     repeats,
     cl2_report_dir,
     cloud_info,
@@ -156,7 +154,7 @@ def collect_clusterloader2(
     else:
         raise Exception(f"No testsuites found in the report! Raw data: {details}")
 
-    _, _, pods_per_node, _ = calculate_config(cpu_per_node, node_count, pods_in_node, provider, service_test, cnp_test, ccnp_test)
+    _, _, pods_per_node, _ = calculate_config(cpu_per_node, node_count, max_pods, provider, service_test, cnp_test, ccnp_test)
     pod_count = node_count * pods_per_node
 
     # TODO: Expose optional parameter to include test details
@@ -219,8 +217,7 @@ def main():
     parser_configure.add_argument("cpu_per_node", type=int, help="CPU per node")
     parser_configure.add_argument("node_count", type=int, help="Number of nodes")
     parser_configure.add_argument("node_per_step", type=int, help="Number of nodes per scaling step")
-    parser_configure.add_argument("max_pods", type=int, help="Maximum number of pods per node")
-    parser_configure.add_argument("pods_in_node", type=int, nargs='?', default=0, help="Number of pods per node")
+    parser_configure.add_argument("max_pods", type=int, nargs='?', default=0, help="Maximum number of pods per node")
     parser_configure.add_argument("repeats", type=int, help="Number of times to repeat the deployment churn")
     parser_configure.add_argument("operation_timeout", type=str, help="Timeout before failing the scale up test")
     parser_configure.add_argument("provider", type=str, help="Cloud provider name")
@@ -256,8 +253,7 @@ def main():
     parser_collect = subparsers.add_parser("collect", help="Collect scale up data")
     parser_collect.add_argument("cpu_per_node", type=int, help="CPU per node")
     parser_collect.add_argument("node_count", type=int, help="Number of nodes")
-    parser_collect.add_argument("max_pods", type=int, help="Maximum number of pods per node")
-    parser_collect.add_argument("pods_in_node", type=int, nargs='?', default=0, help="Number of pods per node")
+    parser_collect.add_argument("max_pods", type=int, nargs='?', default=0, help="Maximum number of pods per node")
     parser_collect.add_argument("repeats", type=int, help="Number of times to repeat the deployment churn")
     parser_collect.add_argument("cl2_report_dir", type=str, help="Path to the CL2 report directory")
     parser_collect.add_argument("cloud_info", type=str, help="Cloud information")
@@ -281,15 +277,15 @@ def main():
 
     if args.command == "configure":
         configure_clusterloader2(args.cpu_per_node, args.node_count, args.node_per_step, args.max_pods,
-                                 args.pods_in_node, args.repeats, args.operation_timeout, args.provider, args.cilium_enabled,
+                                 args.repeats, args.operation_timeout, args.provider, args.cilium_enabled,
                                  args.service_test, args.cnp_test, args.ccnp_test, args.num_cnps, args.num_ccnps, args.dualstack, args.cl2_override_file)
     elif args.command == "validate":
         validate_clusterloader2(args.node_count, args.operation_timeout)
     elif args.command == "execute":
         execute_clusterloader2(args.cl2_image, args.cl2_config_dir, args.cl2_report_dir, args.cl2_config_file,
                                args.kubeconfig, args.provider)
     elif args.command == "collect":
-        collect_clusterloader2(args.cpu_per_node, args.node_count, args.max_pods, args.pods_in_node, args.repeats,
+        collect_clusterloader2(args.cpu_per_node, args.node_count, args.max_pods, args.repeats,
                                args.cl2_report_dir, args.cloud_info, args.run_id, args.run_url,
                                args.service_test, args.cnp_test, args.ccnp_test, args.num_cnps, args.num_ccnps, args.dualstack, args.result_file, args.test_type)
 

pipelines/perf-eval/CNI Benchmark/cnp-ccnp-feature.yml

@@ -24,8 +24,7 @@ stages:
               cpu_per_node: 4
               node_count: $(NODES)
               node_per_step: $(STEP_NODES)
-              max_pods: 110
-              pods_in_node: $(PODS)
+              max_pods: $(MAX_PODS_IN_NODE)
               repeats: 1
               scale_timeout: "15m"
               cilium_enabled: True

steps/engine/clusterloader2/slo/collect.yml

@@ -15,8 +15,11 @@ steps:
 - script: |
     set -eo pipefail
 
+    echo "max : ${MAX_PODS}"
+    echo "${MAX_PODS:-0}"
+
     PYTHONPATH=$PYTHONPATH:$(pwd) python3 $PYTHON_SCRIPT_FILE collect \
-      $CPU_PER_NODE $NODE_COUNT $MAX_PODS ${PODS_IN_NODE:-0} \
+      $CPU_PER_NODE $NODE_COUNT ${MAX_PODS:-0} \
       $REPEATS $CL2_REPORT_DIR "$CLOUD_INFO" $RUN_ID $RUN_URL $SERVICE_TEST ${CNP_TEST:-False} ${CCNP_TEST:-False} ${NUM_CNPS:-0} ${NUM_CCNPS:-0} ${DUALSTACK:-False} $TEST_RESULTS_FILE \
       $TEST_TYPE
   workingDirectory: modules/python/clusterloader2

steps/engine/clusterloader2/slo/execute.yml

@@ -12,8 +12,11 @@ steps:
   - script: |
       set -eo pipefail
 
+      echo "max : ${MAX_PODS}"
+      echo "${MAX_PODS:-0}"
+
       PYTHONPATH=$PYTHONPATH:$(pwd) python3 $PYTHON_SCRIPT_FILE configure \
-        $CPU_PER_NODE $NODE_COUNT $NODE_PER_STEP $MAX_PODS ${PODS_IN_NODE:-0} \
+        $CPU_PER_NODE $NODE_COUNT $NODE_PER_STEP ${MAX_PODS:-0} \
         $REPEATS $SCALE_TIMEOUT $CLOUD $CILIUM_ENABLED \
         $SERVICE_TEST ${CNP_TEST:-False} ${CCNP_TEST:-False} ${NUM_CNPS:-0} ${NUM_CCNPS:-0} ${DUALSTACK:-False} ${CL2_CONFIG_DIR}/overrides.yaml
       PYTHONPATH=$PYTHONPATH:$(pwd) python3 $PYTHON_SCRIPT_FILE execute \