I have been facing this a lot in my current project, and keeping long list of GitHub issues for "tracking" of technical debts hasn't really helped. As these nev

I have been facing this a lot in my current project, and keeping long list of GitHub issues for "tracking" of technical debts hasn't really helped. As these never find its way to our Sprint Backlog. A couple things we started trying to do: (i) Try to estimate more realistically (usually higher). Consider making the feature stable when estimating, drop the "done" and "done done" concepts. A feature is done when it's stable, reviewed and tests green. (ii) Don't accept pull requests that lower the test coverage of our application (at least in modules that we already have coverage). (iii) Pull a few technical debt items to the Sprint Backlog We figured that it was easier to avoid creating more technical debt in rather than trying to retroactively fix all. I've had long discussion with people saying that we should stop for one full week and do a mass cleanup and redesign. Not sure if I fully agree.