Friction report: repo cloning aborted after 1 hour

[bahrmichael]

With this issue I want to report some friction I encountered when asking my local sourcegraph instance to clone the chromium/chromium repo. With 40GB that repo is especially large, and may not be what our customers usually use.

I wanted to clone this repo to replicate an issue with code insights where I need to run queries that turn out super expensive given the codebase.

Since there was a failure (not sure why, probably because I was switching from wifi to ethernet at that point) on the overall list of repositories in sourcegraph, I switched to the individual repo's view. There I could see the progress it was making, and what I think was 1.5-2 hours in, I see a message that the cloning has been aborted because it exceeded the 1 hour threshold. I've only seen this message when I opened the details of the repo being cloned in sourcegraph, and assume that this info is not surfaced when we show at the overall list of repos in my sourcegraph instance.

I saw the message below, and recalled that I can clone a repo and use another command to serve that up to sourcegraph. But since I didn't know about the timeout and found the cloning process easier, I didn't go that route.

After looking into the code I saw that the timeout can be configured via gitLongCommandTimeout, and would suggest to surface the possibility of hitting the timeout before it happens and a user basically has to start over.

Maybe we can track the remaining time, and/or show a note that if the download takes especially large it would be aborted after an hour and that the timeout can be configured via gitLongCommandTimeout?