On July 26th, MovieNight.app went offline for about an hour. Here's why — and why the site is faster because of it.
If you tried to visit MovieNight.app on the morning of July 26th, you probably ran into our maintenance page. First off — thanks for your patience. The site was down for roughly 70 minutes, and it was entirely planned. Here's what we were up to.
What happened
We moved our entire database — every movie, rating, review, list, and versus matchup — to a new, dedicated database server. That's about 26 GB of data and 57 tables, copied, verified row-by-row, and switched over without losing a single record.
We took the site offline during the move for one simple reason: safety. With the app in maintenance mode, nothing could write new data mid-copy, which meant we could verify that every table matched perfectly between the old and new servers before flipping the switch. It did — a 100% match.
Why we did it
MovieNight has been growing, and our movie data pipeline never sleeps — it's constantly pulling in new releases, updated details, streaming availability, and more. Our old managed database was starting to feel like a small kitchen during a dinner rush.
The new server is a dedicated machine tuned specifically for our workload, with significantly more memory dedicated to keeping the hottest data — the movies you're actually browsing — ready to serve instantly.
What it means for you
- Snappier pages. Browsing, search, Top 100 lists, and streaming pages all lean hard on the database — more headroom means faster responses, especially at peak times.
- Fresher movie data. Our update pipeline now has room to breathe, so new releases and changes land in the app faster.
- Room to grow. The new setup gives us the capacity to build bigger features without slowing down what's already here.
- Nightly backups. Your lists, ratings, and rankings are backed up automatically every night.
The nerdy bit
For the curious: the migration was a full dump-and-restore with per-table row-count verification, and the new server's InnoDB buffer pool and redo logs are tuned for our write-heavy movie pipeline. During rehearsal, imports ran at about a third of the speed — tuning the database engine alone made the real cutover 3.5x faster. We even had an automated fallback so the site would have stayed safely on the old database if anything hadn't matched.
Doing better next time
We know an unannounced maintenance page isn't a great experience, even for a planned upgrade. So here's our commitment going forward:
- Zero downtime whenever possible. Most future upgrades will happen live, behind the scenes, with no interruption at all.
- Advance notice when it's not. If we ever do need to take the site offline, we'll announce it ahead of time — right here on the blog — with the date, time, and expected duration.
- A friendlier maintenance page. If you do land on one, it will tell you exactly what's happening and when we'll be back — not just a generic error.
What's next
Nothing you need to do — everything moved with the site, including your account, lists, and rankings. If anything looks off, let us know. Otherwise: enjoy the speed, and go rank some movies.