IPython notebooks are not rendered

I cannot view ipynb files in my private repository. “Sorry, something went wrong. Reload?” But it worked yesterday.

Whereas IPython notebooks from a public repository are rendered properly.

4 Likes

Same here, something must have changed recently so that significant portion of my notebooks no longer render.

The issue is resolved now

Hi! The same problem again. Jupyter notebook does not render.

Error Message: 

Sorry, something went wrong. Reload?

https://github.com/datalanas/Coursera_Capstone/blob/master/Coursera_Capstone.ipynb

2 Likes

Same problem with my repositories. Have you done anything to get it running?

To anyone who works for github:

I have been grappling with the Sorry, something went wrong. [Reload?] issue for months. I have many notebooks in many repositories. (One such example is here).

Whenever a notebook fails with Sorry, something went wrong. [Reload?] , the same notebook will be fine some time later , where “ some time later ” can be anywhere from several minutes to several hours (up to maybe a day). However that doesn’t mean that it won’t start failing again several hours after that.

I am convinced github.com is using a service (hosted somewhere) to render notebooks, and that sometimes this service either gets into a bad state, or crashes; when that happens notebooks won’t render until the service is bounced (or somehow has its “good” state restored).

I would be happy to volunteer my time to help investigate the issue, if it is otherwise not a priority for github.com to allocate the resources. Please let me know if there is anything I can do to help.

This can be very inconvenient, even embarrassing, to say the least, when I am trying to provide support to users of my public repositories and they tell me that the notebooks to which I direct them are not rendering.

Thank you.

2 Likes

This thread (click here) shows that this problem has been going on for at least 4 years and affects quite a large number of users. And the fact that https://nbviewer.jupyter.org/ is a workaround (at least for notebooks that are public on GitHub) is only more evidence that the problem is with GitHub.com.

Does anyone out there know how we can get GitHub’s attention to fix this problem? It seems to me that if they just replace their own notebook rendering software with the Open Source nbviewer then this problem would go away. Anyone who knows coding also knows this is not a difficult problem to solve!

Hey @danielgoldfarb :wave:

Does anyone out there know how we can get GitHub’s attention to fix this problem?

Attention received! In fact, I’ve been looking into these years-old issues alongside our internal dev team to discuss over the past few days.

Right now, I don’t have any immediate updates that I can provide. I have a couple conversations lined up with appropriate folks to break down what’s happened over the years, and what the plans are going forward. My hope is that I can provide a more substantial response inside of a week. However, that’s simply my estimation and not a concrete timeline.

As we discuss, I greatly appreciate your continued patience. :bow:

EDIT: Classic 2020; my expected ETA of a week to provide a more fleshed out response has languished to over a month. I’ve been discussing with multiple parties and once we have an update of substance to provide, I will update this thread with a new response.

1 Like

Since this was bumped in another thread and I’ve replied in-kind over there, I’m going to paste a link to my update, here:

Feel free to read in full, but I think this:

Since I made the comment that I posted in the linked thread, there have been stability improvements made to the hosting platform for our rendering service.

So we should be seeing some of these Sorry, something went wrong. errors become less common in the short term.

Long term? Still up in the air, as I mentioned.

It’s not quite what you’re after, I know. But this is what we have for the time being.

Appreciate everyone’s patience as the work continues!

…is what folks might be interested in for now.

Cheers, everyone.

Any update on this @nethgato?

Python notebooks are becoming the norm more and more in some areas and a lot of people rely on them. Why is this not prioritised?

1 Like

Hi @jorijnsmit :wave:

I caught Daniel’s response before I noticed your ping; apologies! I updated a thread quoted above, here:

The same applies to this conversation as well, and while I’m almost certain the update I’ve provided isn’t what we’re hoping for, it is what we have for now.

Thanks again for keeping this conversation going.

1 Like