Over a few weeks in the month of August, we ran a deployment survey to get a better understanding of where and how Wagtail developers like to deploy their projects. Overall, 153 people responded to the survey and provided us with a snapshot of the community's preferences.
Here are the key results.
Most deployments are to AWS and DigitalOcean
The most common cloud hosting services used by Wagtail developers included Amazon Web Services (AWS) at 24.8% and DigitalOcean at 24.2%. Rounding out the top five were Linode at 13.7%, Heroku with 12.4%, and Google Cloud Platform at 11.8%.
Microsoft Azure (7.2%), Fly.io (8.5%), and Hetzner (which was app. 6.5% after factoring in Hetzner answers listed in the "Other" category) also had strong showings. Ten people (app. 6.5%) who are talented at server wrangling also reported self-hosting their own servers.
Wagtail developers love containers and virtual machines
When asked how they host their Wagtail projects, the majority of respondents (56.9%) said they used containers (like Docker) to host their Wagtail projects. Another 33.3% reported that they use virtual machines (like DigitalOcean's droplets or Heroku's dynos) for hosting their projects. Only 20.9% report using a platform-as-a-service solution like Heroku for hosting their projects and only 4.6% of respondents have experimented with serverless hosting. Around 12.4% of respondents report using an on premises server for hosting rather than cloud hosting.
Postgres is the dominant database, but SQLite is gaining ground
One result that will surprise no one whatsoever is that 90.2% of respondents use Postgres as a database for Wagtail projects. What is a little surprising though is that SQLite came in second with 16.3% of responds saying that they used SQLite as the database for their projects. MySQL at 7.8% and MariaDB at 6.5% have also been used by the survey respondents.
People are struggling with static images and configuring servers
When asked about the deployment steps they struggle with the most, nearly 40% of people reported that they have trouble with the configuration for static images and media files. The second biggest challege for 30.1% of people who responded was configuring the server itself.
Other challenges that people reported included not having enough deployment documentation or examples (22.9%), configuring the Wagtail production server (20.9%), and configuring the database (4.6%). Some challenges reported in the "Others" section included finding affordable hosting services to deploy to as well as wanting more examples of Wagtail best practices.
The community is keen to collaborate with DigitalOcean
When we asked what cloud hosting services people would like to see the Wagtail project partner with, the most popular choice by far at 37.3% was DigitalOcean. AWS came in second at 26.8% with Google Cloud Platform (19%), Fly.io (17.6%), and Vercel (15%) making up the rest of the top 5.
Other assorted comments and suggestions
The open-ended suggestions included requests for more demonstrations of Docker-based deployment, more explicit examples of production cache settings, a demonstration of best practices for security settings, and developing a method for importing content from an older version of Wagtail into a newer version without having to do so many incremental updates.
Many people also left us fun, encouraging comments like "<3", "Keep rocking!", and "Thanks for the great work! 🙏". This is just more evidence that our community is full of truly awesome people. Thank you so much to everyone who left those kind comments and to all the people who gave us their time to fill out the survey.
We definitely plan to keep up the good work and to act on these suggestions as much as we can. If you are willing to share your favorite recipes and settings for deployment, or you are a server nerd who can contribute some expertise or time, please reach out to Meagen Voss through our community Slack or email us at [email protected].
For the purposes of visual clarity, the charts used in this post do not include every response. You can view the raw data from this survey in Google Sheets.