Harnessing Heroku for Efficient App Deployment

In this episode of Running in Production, Dan Bader goes over how he built Real Python from scratch with Django which is a web framework written in Python. The site gets 4 million+ page views a month and it’s hosted on Heroku which costs about $700 USD / month.

What’s really cool about this project is Dan coded the site himself and his deploy process is very by the books which includes performing 1 person code reviews which I found to be great. We covered a lot of ground in this episode! Spoiler alert: he’s using the built-in Django admin.

Topics Include

• 1:47 – The history of Real Python and how Dan ended up taking over
• 3:13 – Replacing a PHP back-end with a custom Django app from scratch
• 3:48 – Dan is the sole developer but he wants that to change soon
• 4:11 – Django has been a happy developer experience and it’s very effective
• 4:47 – Motivation for switching to Django and a break down of Real Python’s features
• 5:46 – Dan tries to keep the code base on the latest stable Django release
• 6:29 – Real Python has been using Python 3.x from day 1
• 7:06 – Staying up to date often takes effort, but it saves time in the long run
• 7:53 – Automated and manual ways to keep your dependencies up to date
• 10:29 – Real Python is mostly a monolithic application by choice
• 12:28 – 19,000+ lines of Python code spread across ~23 Django apps
• 13:44 – A high level break down of the Django apps
• 15:16 – The challenges of trying to break up your apps too early in the dev process
• 18:30 – Server side rendered templates is allowing Dan to be quite efficient
• 19:50 – Django was specifically built to display documents (news articles)
• 20:23 – Real Python uses the built-in Django admin and it’s working great
• 21:48 – Multiple ways to have an admin back-end without an admin dashboard
• 23:21 – The site doesn’t use Django channels or any type of websockets
• 23:51 – Docker is being used locally for development but only partially with Docker Compose
• 24:26 – The Python app is not using Docker mainly due to volume performance issues
• 25:05 – But all local tests are run inside of Docker (even the Python app)
• 26:00 – Dan wishes Docker were faster on MacOS, it would be nice for other developers
• 27:00 – Huey is being used instead of Celery for background tasks
• 28:50 – Huey even supports running periodic scheduled tasks on an interval
• 29:20 – You might need custom code to deal with retries and guarantees
• 30:37 – There’s a lot of responsibility to make sure Real Python stays up
• 31:34 – Dans likes Heroku because it’s easy to get notified of availability issues
• 32:33 – Papertrail provides log access and alerts him over Slack on errors
• 32:48 – For up time monitoring he’s using Freshping
• 32:58 – All errors and notifications get piped to Slack to keep it easy to glance
• 34:16 – What happens if Heroku or AWS goes down? You kind of have to deal with it
• 35:12 – The “up-time value” in building a custom platform / app for your content
• 36:36 – Payments are handled with a reseller called Paddle, they handle taxes and more
• 38:11 – They take a few percent more than Stripe but it’s worth it for Dan
• 38:56 – Dealing with taxes is very complicated and soul draining
• 39:23 – Sometimes these services end up being better on all fronts (even price)
• 40:22 – Paddle supports credit cards, PayPal and Apple Pay through 1 API provided by Paddle
• 41:58 – SendGrid is being used to send out all types of email from the platform
• 43:10 – SendGrid’s batch API endpoint lets you send 1,000 emails in 1 API call
• 44:22 – CloudFlare is being used to convert images to webp (image re-compression)
• 45:20 – CloudFlare (he’s using the paid plan btw) also uses HTTP / 2 in supported browsers
• 45:32 – It also helps with rate limiting because Real Python has been getting attacked
• 45:56 – SSL certificates are handled with CloudFlare and Heroku together
• 46:12 – Real Python is using gunicorn
• 46:25 – How Heroku deals with running and configuring gunicorn
• 48:01 – (4) standard-2x Dynos and (1) standard-1x Worker powers the whole site
• 48:47 – CloudFlare’s CDN helps to offset the dyno’s load by serving all of the static files
• 49:32 – Static files are initially served by WhiteNoise for 1 request then CloudFlare caches it
• 50:59 – A tiny microservice proxy called RuboCrop resizes images to different sizes
• 52:24 – Lots of images and articles require alternate ways of organizing things
• 54:07 – There is a full text search feature built into the site but it might not be enough
• 54:48 – The Heroku add-ons being used are Postgres, Redis, Scheduler and Papertrail
• 55:31 – Hosting costs on Heroku is about $700 / month (including a staging environment)
• 56:36 – It’s totally worth it for Dan, the value he gets is massive
• 57:32 – As a solo developer you need to think about the time vs money vs sanity trade off
• 59:03 – Walking though a code deploy from development to staging to production
• 1:00:36 – Heroku’s metrics dashboard makes it easy to rollback if things look off
• 1:01:52 – 1 person code reviews are awesome and it allows for scaling up devs later
• 1:02:30 – Heroku handles database backups and user uploaded files are sent to S3
• 1:03:45 – Dan thinks he can rebuild everything very quickly if disaster struck
• 1:04:18 – Vimeo is being used to host videos and provide the video player
• 1:05:37 – Best tips? Embrace the golden age of the solo web developer being viable
• 1:07:00 – It’s so easy to initially get caught up in building something super difficult
• 1:08:27 – One of Dan’s biggest mistakes at a previous company was going microservices
• 1:10:26 – Before it was released Real Python ran on WordPress but Dan decided against it
• 1:11:19 – Philosophically it kind of sucks to have a big Python site running on PHP
• 1:11:55 – Building your own tutorial platform helps you come up with tutorial content
• 1:12:34 – Find Dan on Twitter @dbader_org and his personal site. You can also find Real Python on Twitter, YouTube and Instagram as well as realpython.com

Links

📄 References

• https://www.kickstarter.com/projects/721054906/real-python-advanced-web-development-featuring-dja
• https://www.amazon.com/Python-Tricks-Buffet-Awesome-Features-ebook/dp/B0785Q7GSY
• https://en.wikipedia.org/wiki/Waterfall_model
• https://en.wikipedia.org/wiki/Lawrence_Journal-World
• https://realpython.com/python-virtual-environments-a-primer/
• https://en.wikipedia.org/wiki/WebP
• https://developer.mozilla.org/en-US/docs/Learn/HTML/Multimedia_and_embedding/Responsive_images

⚙️ Tech Stack

• django →
• python →
• circle-ci →
• cloudflare →
• docker →
• freshping →
• heroku →
• paddle →
• papertrail →
• postgres →
• redis →
• s3 →
• sendgrid →
• slack →
• vimeo →

🛠 Libraries Used

• https://github.com/celery/celery
• https://github.com/coleifer/huey
• https://gunicorn.org/
• https://github.com/evansd/whitenoise

Support the Show

This episode does not have a sponsor and this podcast is a labor of love. If you want to support the show, the best way to do it is to purchase one of my courses or suggest one to a friend.

• Dive into Docker is a video course that takes you from not knowing what Docker is to being able to confidently use Docker and Docker Compose for your own apps. Long gone are the days of "but it works on my machine!". A bunch of follow along labs are included.

• Build a SAAS App with Flask is a video course where we build a real world SAAS app that accepts payments, has a custom admin, includes high test coverage and goes over how to implement and apply 50+ common web app features. There's over 20+ hours of video.