BUG: traceback in the parallel notebook doesn't fail CI status

[bsipocz]

There is a traceback in the Dask case for the parallel notebook, and somehow this doesn't fail the CI or execution and rendering.

I see no sign of this in the pytest jobs (which uses nbval), but the traceback is present with myst-nb execution.

https://github.com/Caltech-IPAC/irsa-tutorials/actions/runs/10466243659/job/28982790720#step:5:242

https://caltech-ipac.github.io/irsa-tutorials/tutorials/parallelize/Parallelize_Convolution.html#process-using-dask

start = time.time()
image = np.zeros((3000, 3000))
for _ in range(100):
    for j in range(num_cpus):
        big_future = client.scatter((image, filters[j % num_cpus]))
        future = client.submit(fconv, big_future)
duration_dask = time.time() - start
print("Dask duration = {:.1f}, speedup = {:.2f}"
      .format(duration_dask, duration_conv*num_cpus / duration_dask))

2024-08-20 06:26:59,687 - distributed.worker - ERROR - Compute Failed
Key:       fconv-62d3fb08cdcd4722add20d9c193727a4
State:     executing
Function:  fconv
args:      ((array([[0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       ...,
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.],
       [0., 0., 0., ..., 0., 0., 0.]]), array([[ 0.04511129, -0.36965459, -0.95263628, -0.14601954],
       [-0.59376302, -0.84903493, -0.89073568, -0.66741073],
       [-0.47099116, -1.10903786,  1.44708216,  0.12989074],
       [-1.48281771, -1.25057209, -0.32882889,  0.46543018]])))
kwargs:    {}
Exception: 'TypeError("fconv() missing 1 required positional argument: \'random_filter\'")'
Traceback: ''