We’ve been busy shipping improvements across the platform. Here’s a summary of what’s new this month.
Scale applications with replicas (up to 10)
You can now run multiple instances of any application (web or worker), up to 10 replicas, straight from Settings → Replicas.
Load is spread across instances, which helps when you hit spikes, timeouts, or CPU pegged at 100%. Rolling deploys are less painful too: you’re not betting everything on a single box.
Try it: open an app → Settings → Replicas.
Auto deploy for GitHub repos
Auto deploy lives under Build and deploy settings.
Flip it on, and every push to your selected branch triggers a build + deploy of the latest commit. No more manual “Did I deploy main yet?” checks.
Right now, this is GitHub only. Other Git providers still deploy the way you’re used to: pick a commit or use Deploy latest commit (below).
Deploy latest commit (one click)
From the service header menu:
- Deploy latest commit — ship whatever’s at the tip of the branch
- Deploy specific commit — when you need a particular SHA
Small thing, but it saves a lot of clicking when you’re iterating fast.
Environment variables: edit in bulk
Env vars now work like you’d expect in a proper config UI:
- Add, edit, or remove as many as you want
- Review everything
- Apply or Discard in one shot
Database metrics: connections chart
On managed databases, Metrics now includes a Connections chart alongside storage.
Handy when you’re debugging pool sizing, connection leaks, or “why is my app opening 30 connections to Postgres?”
Dashboard QoL
We’ve also shipped a pile of quality-of-life work. The dashboard should feel snappier overall: faster loading, skeleton states while lists hydrate, and less jank when you’re jumping between services.