Hey folks, thanks again for your patience over the last few weeks while we update our platform and infrastructure (which we previously mentioned). We’ve rolled out the changes to the project container to rectify the (brief but too long) Python downgrade and we think the container is once again back to its former glory and the recent project and fleet disruptions are behind us.
These changes and the disruption they’ve caused are steps in the process of improving project fleet management and flexibility. We’re working towards being able to keep the project container more up-to-date and on fleet efficiency with an eye towards being able to test changes to the container and the fleet with significantly less chaos, but this requires “testing” some of these changes in our production environment - many of the issues we’ve recently uncovered are only perceptible at production scale so the pre-production testing we do as a matter of course doesn’t catch all of the flaws already in the system or ones introduced by recent changes.
That said, we’ve learned a lot about what won’t work well enough over the last few weeks, and what we’ve learned will help us make the next rounds of changes more quickly (we’ve been working on some of the upcoming container changes for more than a year) and safely.
There’s more work for us to do before we can adopt a manageable cadence for updates like this but our goals are to be able to provide you with a more-stable project fleet, a regularly-updated project environment, and predictable maintenance windows. There’s a bit more turbulence ahead while we close in on those objectives and we’ll keep you updated here and on https://status.glitch.com in case of incidents.
Thanks again for sticking with us through this - we’re excited for things to come and can’t wait to put this boring-but-necessary work behind us and get back to building safe and engaging web experiences on our favorite place to code!