Heroku is very fortunate to have a strong community of developers that are excited and passionate about our product. Every day we hear from customers who tell us how much easier Heroku has made their lives, and they frequently share stories about interesting technical projects we've helped them bring to life.
Our customers love us, and we love them right back. Today we'll take a look at a few blog posts and applications from Heroku users that illustrate what makes our community so special. We hope you enjoy the tour. If you have Heroku stories of your own you'd like to share, we'd love to hear them!
This article comes to us from Yoni Weisbrod, a JS & React Native developer from Ivy. Ivy makes a community and business management tool for interior designers. They ran into some challenges when their worker dynos
maxing out their memory while generating some particularly hefty internal reports.
Yoni and team decided to create a couple of new workers named Julia and Winston (after the characters from 1984). They use their new workers to dynamically handle the increased load from their report generation tasks. Julia uses the Heroku API to spin up Winston whenever a new report needs to be generated, then adds the report job to Winston's queue. Winston generates the report on demand and spins back down when the mischief has been managed.
The beauty of this solution is that Winston is only used for the exact period of time it takes to generate the report, and since Heroku prorates dyno usage to the second, Ivy only ends up paying for the time they need to get their reports.
Imagine you're chatting with a friend on Telegram and you're reminded of a hilarious cat video you want to share. Think of all the tedium you'll have to endure while you switch applications, search for the video, get a shareable link, switch back to Telegram, and finally paste the link. This is the 21st century! We shouldn't have to find our own videos. This sounds like the perfect use case for a bot.
You can read all about building your own basic greeting bot in this article from Roman Gaponov, CEO and Co-Founder of the DjangoStars agile software development company from Kiev, Ukraine. Roman uses Python on Heroku with the Telegram Bots API to build a simple bot that replies with a distinct message based on the time of day.
If you're interested in building the YouTube lookup bot described above, you'll be much closer after you've walked through the development process described in Roman's article. Have a look at the YouTube Data API to get the rest of the way there.
The Werwolf Game (also known as Mafia) was originally created by Dmitry Davidoff. It's a popular game among developer communities because it requires minimal setup and offers a unique experience every time you play it.
Gabriela's beautiful site brings Werewolf to your online friends as well, with a shared deck online that allows you and your frenemies to play remotely over Skype or any other voice medium. Now you have no excuses not to bond remotely over terror under a full moon.
Eleanor Stribling brings us a well-written post detailing PostgreSQL setup for Django applications. When you first set up a new Django app, it uses SQLite by default. Because Heroku uses an ephemeral file system (and SQLite is a disk-backed database), it's very likely that you would lose data if you did manage to get SQLite to work on Heroku.
Heroku is pretty good at warning you when we detect that you're attempting to use SQLite, so you're not likely to actually lose any data. If you'd like some more information about SQLite on Heroku, check out this devcenter article.