[configure] Localising pod start-up latency to scheduling, image pull, volume attach, or CNI

Author: jing2uo

docs/en/solutions/Localising_pod_start_up_latency_to_scheduling_image_pull_volume_attach_or_CNI.md

@@ -113,7 +113,7 @@ When the slowness is intermittent, run the same extraction across a recent windo
 
    ```bash
    kubectl debug node/<node> -it --profile=sysadmin --image=<utility-image> \
-     -- chroot /host crictl pull <same-image>
+     -- crictl pull <same-image>
    ```
 
    A consistent multi-second pull on a known-good image points at registry network or rate-limiting; a fast pull while the original pod is slow points at parallel-pull saturation on the kubelet.