fix(tools): changes to experiment infra to use custom precompute engine

asap-tools/experiments/experiment_run_e2e.py

@@ -518,6 +518,12 @@ def main(cfg: DictConfig):
                     http_port=http_port,
                     remote_write_port=args.remote_write_base_port,
                 )
+                # For precompute mode the query engine IS the Prometheus remote-write
+                # target (port remote_write_base_port). Prometheus is already running
+                # and retrying writes against that port, so block until the query
+                # engine's HTTP server is accepting connections before proceeding.
+                if args.streaming_engine == "precompute":
+                    query_engine_service.wait_until_ready()
 
         # Start system exporters (node_exporter, blackbox_exporter, cadvisor)
         system_exporters_service.start(cfg.experiment_params)

asap-tools/experiments/experiment_utils/services/query_engine.py

@@ -524,16 +524,17 @@ def _is_healthy_bare_metal(self) -> bool:
     def _is_healthy_containerized(self) -> bool:
         """Check if Rust QueryEngine is healthy using containerized deployment."""
         try:
-            # Check if container is running
-            result = subprocess.run(
-                ["docker", "inspect", "-f", "{{.State.Running}}", self.container_name],
-                capture_output=True,
-                text=True,
-                check=True,
+            cmd = f"docker inspect -f '{{{{.State.Running}}}}' {self.container_name}"
+            result = self.provider.execute_command(
+                node_idx=self.node_offset,
+                cmd=cmd,
+                cmd_dir=None,
+                nohup=False,
+                popen=False,
+                ignore_errors=True,
             )
+            assert isinstance(result, subprocess.CompletedProcess)
             return result.stdout.strip() == "true"
-        except subprocess.CalledProcessError:
-            return False
         except Exception:
             return False