feat: support granular aggressive mode config in scheduled rebalancing

Author: vladklokun

castai/data_source_resource_rebalancing_schedule_test.go

@@ -50,7 +50,10 @@ func TestRebalancingScheduleDataSourceRead(t *testing.T) {
                     "evictGracefully": false,
                     "aggressiveMode": false,
                     "aggressiveModeConfig": {
-                        "ignoreLocalPersistentVolumes": true	
+                        "ignoreLocalPersistentVolumes": true,
+						"ignoreProblemJobPods": true,
+						"ignoreProblemRemovalDisabledPods": true,
+						"ignoreProblemPodsWithoutController": true
                     }
                 },
                 "numTargetedNodes": 20,
@@ -131,6 +134,9 @@ launch_configuration.# = 1
 launch_configuration.0.aggressive_mode = false
 launch_configuration.0.aggressive_mode_config.# = 1
 launch_configuration.0.aggressive_mode_config.0.ignore_local_persistent_volumes = true
+launch_configuration.0.aggressive_mode_config.0.ignore_problem_job_pods = true
+launch_configuration.0.aggressive_mode_config.0.ignore_problem_pods_without_controller = true
+launch_configuration.0.aggressive_mode_config.0.ignore_problem_removal_disabled_pods = true
 launch_configuration.0.execution_conditions.# = 1
 launch_configuration.0.execution_conditions.0.achieved_savings_percentage = 15
 launch_configuration.0.execution_conditions.0.enabled = true

castai/resource_rebalancing_schedule.go

@@ -131,7 +131,22 @@ func resourceRebalancingSchedule() *schema.Resource {
 									"ignore_local_persistent_volumes": {
 										Type:        schema.TypeBool,
 										Required:    true,
-										Description: "Rebalance workloads using local-path Persistent Volumes. THIS WILL RESULT IN DATA LOSS.",
+										Description: "Rebalance workloads that use local-path Persistent Volumes. THIS WILL RESULT IN DATA LOSS.",
+									},
+									"ignore_problem_job_pods": {
+										Type:        schema.TypeBool,
+										Required:    true,
+										Description: "Pods spawned by Jobs or CronJobs will not prevent the Rebalancer from deleting a node on which they run. WARNING: When true, pods spawned by Jobs or CronJobs will be terminated if the Rebalancer picks a node that runs them. As such, they are likely to lose their progress.",
+									},
+									"ignore_problem_removal_disabled_pods": {
+										Type:        schema.TypeBool,
+										Required:    true,
+										Description: `Pods that are marked with "removal disabled" will not prevent the Rebalancer from deleting a node on which they run. WARNING: When true, such pods will be evicted and disrupted.`,
+									},
+									"ignore_problem_pods_without_controller": {
+										Type:        schema.TypeBool,
+										Required:    true,
+										Description: "Pods that don't have a controller (bare pods) will not prevent the Rebalancer from deleting a node on which they run. WARNING: When true, such pods might not restart, since they have no controller to do it.",
 									},
 								},
 							},
@@ -314,7 +329,10 @@ func stateToSchedule(d *schema.ResourceData) (*sdk.ScheduledrebalancingV1Rebalan
 		aggressiveModeConfigSection := launchConfigurationData["aggressive_mode_config"].([]any)
 		if len(aggressiveModeConfigSection) != 0 {
 			aggressiveModeConfig = &sdk.ScheduledrebalancingV1AggressiveModeConfig{
-				IgnoreLocalPersistentVolumes: lo.ToPtr(aggressiveModeConfigSection[0].(map[string]any)["ignore_local_persistent_volumes"].(bool)),
+				IgnoreLocalPersistentVolumes:       lo.ToPtr(aggressiveModeConfigSection[0].(map[string]any)["ignore_local_persistent_volumes"].(bool)),
+				IgnoreProblemJobPods:               lo.ToPtr(aggressiveModeConfigSection[0].(map[string]any)["ignore_problem_job_pods"].(bool)),
+				IgnoreProblemRemovalDisabledPods:   lo.ToPtr(aggressiveModeConfigSection[0].(map[string]any)["ignore_problem_removal_disabled_pods"].(bool)),
+				IgnoreProblemPodsWithoutController: lo.ToPtr(aggressiveModeConfigSection[0].(map[string]any)["ignore_problem_pods_without_controller"].(bool)),
 			}
 		}
 
@@ -363,7 +381,10 @@ func scheduleToState(schedule *sdk.ScheduledrebalancingV1RebalancingSchedule, d
 		if schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig != nil {
 			launchConfig["aggressive_mode_config"] = []map[string]any{
 				{
-					"ignore_local_persistent_volumes": schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig.IgnoreLocalPersistentVolumes,
+					"ignore_local_persistent_volumes":        schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig.IgnoreLocalPersistentVolumes,
+					"ignore_problem_job_pods":                schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig.IgnoreProblemJobPods,
+					"ignore_problem_removal_disabled_pods":   schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig.IgnoreProblemRemovalDisabledPods,
+					"ignore_problem_pods_without_controller": schedule.LaunchConfiguration.RebalancingOptions.AggressiveModeConfig.IgnoreProblemPodsWithoutController,
 				},
 			}
 		}

examples/gke/gke_cluster_autoscaler_policies/castai.tf

@@ -26,7 +26,7 @@ module "castai-gke-iam" {
 }
 
 module "castai-gke-cluster" {
-  source = "castai/gke-cluster/castai"
+  source = "../../../../terraform-castai-gke-cluster"
 
   api_url                = var.castai_api_url
   castai_api_token       = var.castai_api_token