Improved error handling

api/app.py

@@ -439,12 +439,33 @@ def _wait_abort_then_destroy(attack_range_id: str, config_path: str) -> None:
         pass
 
 
+def _set_terraform_running(attack_range_id: str, running: bool) -> None:
+    """Track whether Terraform is currently provisioning infrastructure for a build."""
+    with operations_lock:
+        op = running_operations.get(attack_range_id)
+        if op is not None:
+            op["terraform_running"] = running
+
+
+def _is_terraform_running(attack_range_id: str) -> bool:
+    with operations_lock:
+        op = running_operations.get(attack_range_id)
+        return bool(op and op.get("terraform_running"))
+
+
+def _is_abort_allowed(attack_range_id: str, status: str) -> bool:
+    build_statuses = ("queued", "build_vpn", "build_lab")
+    return status in build_statuses and not _is_terraform_running(attack_range_id)
+
+
 def _check_abort_and_set_aborted(attack_range_id: str, config_path: Optional[str] = None) -> bool:
     """If abort_requested is set for this attack_range_id, set status to aborted and return True. Else return False."""
     with operations_lock:
         op = running_operations.get(attack_range_id)
         if not op or not op.get("abort_requested"):
             return False
+        if op.get("terraform_running"):
+            return False
         running_operations[attack_range_id]["status"] = "aborted"
         running_operations[attack_range_id]["end_time"] = datetime.now().isoformat()
         path = config_path or op.get("config_path") or get_config_path_from_attack_range_id(attack_range_id)
@@ -474,7 +495,11 @@ def run_build_vpn_phase(config: Dict[str, Any], config_path: str, attack_range_i
         controller = AttackRangeController(config, config_path=config_path)
 
         # Build VPN phase (handles all steps including status updates; checks abort between steps)
-        router_public_ip, wireguard_config = controller.build_vpn_phase(attack_range_id, abort_check=lambda: _check_abort_and_set_aborted(attack_range_id, config_path))
+        router_public_ip, wireguard_config = controller.build_vpn_phase(
+            attack_range_id,
+            abort_check=lambda: _check_abort_and_set_aborted(attack_range_id, config_path),
+            terraform_running_callback=lambda running: _set_terraform_running(attack_range_id, running),
+        )
 
         wireguard_config_path = os.path.join(WIREGUARD_CONFIG_DIR, f"{attack_range_id}.conf")
 
@@ -822,6 +847,11 @@ def destroy_attack_range(body: DestroyRequest):
             status = config_for_status.get("general", {}).get("status") or ""
 
         if status in build_statuses:
+            if _is_terraform_running(attack_range_id):
+                return jsonify(ErrorResponse(
+                    message="Cannot destroy while Terraform is provisioning infrastructure",
+                    details="Wait for Terraform to finish, then retry destroy or abort"
+                ).model_dump()), 400
             with operations_lock:
                 op = running_operations.get(attack_range_id)
                 if op:
@@ -886,7 +916,7 @@ def destroy_attack_range(body: DestroyRequest):
     tags=[attack_range_tag],
     responses={200: DestroyResponse, 400: ErrorResponse, 404: ErrorResponse, 500: ErrorResponse},
     summary="Abort attack range build",
-    description="Abort a build operation in progress. Sets status to 'aborted'."
+    description="Abort a build operation in progress. Sets status to 'aborted'. Not allowed while Terraform is provisioning infrastructure."
 )
 def abort_attack_range(body: DestroyRequest):
     """Abort a build operation by setting abort_requested flag and status to aborted."""
@@ -921,6 +951,12 @@ def abort_attack_range(body: DestroyRequest):
                 details="Abort can only be called during build (queued, build_vpn, build_lab)"
             ).model_dump()), 400
 
+        if _is_terraform_running(attack_range_id):
+            return jsonify(ErrorResponse(
+                message="Cannot abort while Terraform is provisioning infrastructure",
+                details="Abort is disabled during Terraform init/apply to avoid leaving broken cloud resources. Retry after provisioning completes."
+            ).model_dump()), 400
+
         # Set abort_requested flag
         with operations_lock:
             op = running_operations.get(attack_range_id)
@@ -1025,6 +1061,9 @@ def get_attack_range_status(path: AttackRangeIdPath):
         if "result" in operation and "config_file" not in operation["result"]:
             operation["result"]["config_file"] = config_path
 
+    build_statuses = ("queued", "build_vpn", "build_lab")
+    operation["abort_allowed"] = operation.get("status") in build_statuses and not operation.get("terraform_running", False)
+
     # Validate and return as OperationStatusResponse
     return jsonify(OperationStatusResponse(**operation).model_dump()), 200
 

api/models.py

@@ -84,6 +84,8 @@ class OperationStatusResponse(BaseModel):
     error: Optional[str] = Field(None, description="Error message (for failed operations)")
     error_phase: Optional[str] = Field(None, description="Build phase where error occurred")
     traceback: Optional[str] = Field(None, description="Error traceback (for failed operations)")
+    terraform_running: Optional[bool] = Field(None, description="True while Terraform init/apply is in progress")
+    abort_allowed: Optional[bool] = Field(None, description="True when abort is permitted for the current build state")
 
 
 class ServerInfo(BaseModel):

app/src/pages/attack-ranges/[id].astro

@@ -58,8 +58,13 @@ if (id) {
 				<div class="basic-info-header">
 					<h3>Basic Information</h3>
 					{['queued', 'build_vpn', 'build_lab'].includes(attackRange.status) ? (
-						<button class="btn-warning abort-button" data-id={attackRange.attack_range_id}>
-							Abort
+						<button
+							class="btn-warning abort-button"
+							data-id={attackRange.attack_range_id}
+							disabled={attackRange.terraform_running}
+							title={attackRange.terraform_running ? 'Abort is not allowed while Terraform is provisioning infrastructure' : undefined}
+						>
+							{attackRange.terraform_running ? 'Provisioning...' : 'Abort'}
 						</button>
 					) : (
 						<button
@@ -626,12 +631,14 @@ if (id) {
 			abortButton.addEventListener('click', async () => {
 				const attackRangeId = (abortButton as HTMLElement).dataset.id;
 				if (!attackRangeId) return;
+
+				const btn = abortButton as HTMLButtonElement;
+				if (btn.disabled) return;
 
 				if (!confirm(`Are you sure you want to abort the build for attack range ${attackRangeId}?`)) {
 					return;
 				}
 
-				const btn = abortButton as HTMLButtonElement;
 				btn.disabled = true;
 				btn.textContent = 'Aborting...';
 
@@ -954,13 +961,28 @@ if (id) {
 		const attackRangeId = window.location.pathname.split('/').pop();
 		if (attackRangeId) {
 			let lastStatus: string | null = null;
+			let lastTerraformRunning: boolean | null = null;
 			let pollingInterval: number | null = null;
 
 			// Get initial status from the page
 			const statusElement = document.querySelector('.status-badge');
 			if (statusElement) {
 				lastStatus = statusElement.textContent?.trim() || null;
 			}
+			const initialAbortButton = document.querySelector('.abort-button') as HTMLButtonElement | null;
+			if (initialAbortButton) {
+				lastTerraformRunning = initialAbortButton.disabled;
+			}
+
+			const updateAbortButton = (terraformRunning: boolean) => {
+				const btn = document.querySelector('.abort-button') as HTMLButtonElement | null;
+				if (!btn) return;
+				btn.disabled = terraformRunning;
+				btn.textContent = terraformRunning ? 'Provisioning...' : 'Abort';
+				btn.title = terraformRunning
+					? 'Abort is not allowed while Terraform is provisioning infrastructure'
+					: '';
+			};
 
 			// Only poll if status is not in a final state
 			const finalStates = ['running', 'error'];
@@ -973,6 +995,12 @@ if (id) {
 						if (response.ok) {
 							const data = await response.json();
 							const currentStatus = data.status;
+							const terraformRunning = Boolean(data.terraform_running);
+
+							if (terraformRunning !== lastTerraformRunning) {
+								updateAbortButton(terraformRunning);
+								lastTerraformRunning = terraformRunning;
+							}
 
 							// Only reload if status changed
 							if (currentStatus !== lastStatus) {

app/src/utils/api-server.ts

@@ -54,6 +54,8 @@ export interface AttackRange {
 	error?: string;
 	error_phase?: string;
 	traceback?: string;
+	terraform_running?: boolean;
+	abort_allowed?: boolean;
 }
 
 export async function fetchTemplates(): Promise<{ templates: Template[]; providerAvailability: Record<string, ProviderAvailability> }> {

attack_range/attack_range_controller.py

@@ -228,12 +228,32 @@ def build(self, two_phase: bool = False) -> None:
             self.logger.info(f"Configuration saved to: {config_file_path}")
             self.logger.info("="*80 + "\n")
 
-    def build_vpn_phase(self, attack_range_id: str, abort_check: Optional[Callable[[], bool]] = None) -> tuple:
+    def _run_terraform_step(
+        self,
+        terraform_running_callback: Optional[Callable[[bool], None]],
+        step: Callable[[], None],
+    ) -> None:
+        """Run a Terraform step and report running state via callback."""
+        if terraform_running_callback:
+            terraform_running_callback(True)
+        try:
+            step()
+        finally:
+            if terraform_running_callback:
+                terraform_running_callback(False)
+
+    def build_vpn_phase(
+        self,
+        attack_range_id: str,
+        abort_check: Optional[Callable[[], bool]] = None,
+        terraform_running_callback: Optional[Callable[[bool], None]] = None,
+    ) -> tuple:
         """
         Build VPN infrastructure (Phase 1).
 
         :param attack_range_id: Attack range ID
         :param abort_check: Optional callable(); if it returns True, build is aborted (raises RuntimeError).
+        :param terraform_running_callback: Optional callable(bool); called with True before and False after Terraform operations.
         :return: Tuple of (router_public_ip, wireguard_config)
         """
         if abort_check and abort_check():
@@ -268,12 +288,18 @@ def build_vpn_phase(self, attack_range_id: str, abort_check: Optional[Callable[[
             raise RuntimeError("Build aborted")
 
         # Initialize terraform
-        self.terraform_manager.init(backend_was_created)
+        self._run_terraform_step(
+            terraform_running_callback,
+            lambda: self.terraform_manager.init(backend_was_created),
+        )
         if abort_check and abort_check():
             raise RuntimeError("Build aborted")
 
         # Apply terraform
-        self.terraform_manager.apply()
+        self._run_terraform_step(
+            terraform_running_callback,
+            lambda: self.terraform_manager.apply(),
+        )
         if abort_check and abort_check():
             raise RuntimeError("Build aborted")