//   * size(batching_dimensions) <= 3 (TFLite support restriction)
//   * size(contracting_dimensions) = 1
//   * Input tensors are per-tensor uniform quantized (i8->f32)
//     tensors (full integer) with shape [..., r_x, c_x] or [..., c_x, r_x].
//   * The filter tensor is a per-tensor uniform quantized (i8->f32) tensor
//     (constant or activation) with shape [..., r_y, c_y] or [..., c_y, r_y].

	// Use WithoutCancel instead of a new context.TODO() to propagate trace context
	// Call the container runtime's SyncPod callback
	sctx := context.WithoutCancel(ctx)
	result := kl.containerRuntime.SyncPod(sctx, pod, podStatus, pullSecrets, kl.backOff)
	kl.reasonCache.Update(pod.UID, result)
	if err := result.Error(); err != nil {
		// Do not return error if the only failures were pods in backoff