Deploy for large site times out

I’m working on a large site that has almost 30k files in the built site, and this number is only going to increase. Since I’m currently actively developing the site, I often have builds where a large majority of the built files are modified since I’m working on template files. This means that my CI server’s jobs often time out at an hour when trying to deploy the finished build to Netlify.

I had hoped that I could upload a zip of the complete build from my CI server and Netlify could then uncompress the build and discard the unchanged files, but it doesn’t seem this is an option. Or have I missed it?

I understand that there may be settings that can be turned on for my account that would help me reduce the upload time substantially. Will these settings help in my situation? And, if so, how can I get these settings changed?

You can use a zip file as described here:

…but we will still have to compare every single file’s checksums with ones already in our system, so it’s not clear it will go any faster. Worth a try though!

Your long term solution is not to make such sweeping changes, or to make them in a local copy or branch, and deploy your files in batches to some other site (or on some non-production branch at Netlify), and then submit a deploy containing no new files, to make sure you don’t hit that timeout.

Hard to say if the settings we can set would help without a link to your deploy logs for a failed deploy - mind sharing those? Then we can confirm whether a thing we can configure was involved :slight_smile:

1 Like

Unfortunately these deploys haven’t appeared in the list of deploys (production or previews) in the Netlify dashboard. Why would that be the case? I did stop builds for the site to disable the automatic git deploys, but I thought I would still be able to deploy via the API and CLI.

Ah, if they fail before they complete it may not leave logs here. I had missed the detail that you were running on “your CI server”!

When the job at your CI provider fails - how does it fail? Just “you can’t run for longer than an hour there”? In that case you’ll need to follow my advice from the comment yesterday. Or instead some error from our side during the deploy sending? Showing that error in detail would be helpful of that is the case.