Mitigate data loss in web forms when a request fails
Description
I have recently begun collaborating with someone in a new GitLab project. The other day we had our first meeting with a potential customer, during which he took copious notes in the gitlab issue page. When he clicked "save changes" the page updated to the issues list as if nothing had happened, and all of the notes were lost. We tried pressing the back button in the hopes that we could recover the notes but to no avail.
I personally would never take valuable notes in a web form because I don't really trust web apps in general. My colleague is less technical, and was furious. He no longer wants to use GitLab and I am now faced with a difficult decision about which platform to recommend. To paraphrase him: "It's insane that in 2018 an app like this does not have autosave." At first I thought this was naive of him, but upon further reflection I think he's right. That my expectations of web apps are so low is a weak excuse.
GitLab should know how badly this sort of detail can affect potential customers. If there is a way that you can build in some resiliency so that a request failure preserves client-side content, then I imagine it will be well worth the investment. Even if the page just puked an error message with the text contents somewhere in the dump that would have been a tenth as bad as the silent failure and total data loss that we experienced.
Thank you for all of your hard work--on the whole GitLab is a great product. I hope this proves helpful.