Feature Requests

I propose integrating the UV tool into the Python project deployment workflow to significantly reduce installation time for projects with extensive requirements.txt dependencies. Why UV? Extremely fast: UV, written in Rust, dramatically speeds up dependency resolution and installation. Better performance: It outperforms traditional package managers like pip in handling large dependency lists. Efficient caching: Reduces redundant downloads and speeds up subsequent installs. Use Case For projects with many dependencies, deploying via pip can be slow. Replacing pip install -r requirements.txt with uv pip install -r requirements.txt can optimize installation speed, improving CI/CD efficiency and local setup times. Proposal Evaluate UV’s performance against pip for large-scale Python projects. Provide compatibility checks for existing dependency resolution workflows. Consider adding UV as a recommended installation method in deployment guides. UV Repo: https://github.com/astral-sh/uv This enhancement could streamline Python project deployments, making development and production workflows more efficient.

To serve a static page, we need to manually generate the site and create a Docker image with nginx to serve it. This could be directly supported by Koyeb where we just provide a GitHub repo. Site could be built using buildpack and served by some global Nginx.

We do not support backups of our Serverless Databases. We want to provide this feature to ensure data loss prevention, and historical data retention

Add support for inbound TCP connections/public IPs for supporting different protocols like RTMP.

Add support for inbound UDP connections/public IPs for supporting different protocols like DNS.

To reduce costs on applications which do not have constant traffic and are not latency sensitive, a scale-to-zero feature would be useful.

It would be good to have an option to add managed database to the App. For example, I see that it is probably possible to add MariaDB Galera cluster with proxysql. This may require having attachable storage, but for a small DB even the container storage should be sufficient while maintaining synchronization across members of the cluster.

Our current log system displays logs from the last 72 hours without any option to filter or search through the logs. It makes it hard to troubleshoot if the service generates a lot of logs. We want to enhance the experience with logs and enrich it with filter and search capabilities to ease troubleshooting.

Would be nice to support cron and rate scheduling