Comparison
Platform vs Legacy PaaS
The first generation of PaaS made deploys simple. The platform keeps that, and adds what growing teams outgrew it for.
TL;DR
- Keep push-to-deploy, and add private networking and preview environments.
- Per-second billing for instances, so idle services stop costing a fixed monthly fee.
- Choose a region for every service and database, with no add-on marketplace to wire up.
Side by side
Familiar workflow, fewer limits
Where an older PaaS stops, and what you get instead.
| Capability | Platform | Legacy PaaS |
|---|---|---|
| Deploys | ||
| Deploy on push | Included | Included |
| Preview environments | Included | Paid add-on |
| Zero-downtime deploys | Included | Higher tiers only |
| Background jobs and cron jobs | Built in | Extra processes |
| Infrastructure | ||
| Private networking | Included | Not included |
| Regions | Choose per service | Two or three |
| Managed Postgres | Built in | Marketplace add-on |
| Autoscaling | Included | Manual or extra cost |
| Pricing | ||
| Billing | Per instance, per second | Fixed per process, per month |
| Free static sites | Included | Not included |
| Instances sleep on free plans | Not included | Included |
Why teams switch
The reasons we hear most from teams that outgrew their first PaaS.
Pay for what runs
Per-second billing means quiet services cost almost nothing overnight.
Private by default
Services and databases talk over a private network instead of the public internet.
Previews included
Every pull request gets a preview environment on every paid plan.
Regions you choose
Run each service close to its users and keep data where it needs to stay.
Move over without a rewrite
Your build already works. The move is mostly configuration.
- 01Cloning repository from main
- 02Detected Ruby runtime
- 03Procfile found: web, worker
- 04Build completed in 51s
- 05Health check passed
- 06Deploy is live
- Web service
- Worker
- Cron job
- Postgres
- Cache
- Static site
- Database restored from a dump
- Environment variables imported
- Custom domain verified
- TLS certificate issued
- Traffic moved to the platform
- Old apps scaled to zero
“We kept the workflow everyone liked and dropped the add-on bill. Preview environments alone changed how we review work.”
01Does my Procfile still work?
Yes. Each process type becomes a service: web becomes a web service, and other entries become workers or cron jobs.
02How do I move my Postgres data?
Create a managed Postgres database, then restore a dump into it. Larger databases can replicate first and switch over with a few seconds of read-only time.
03Will my costs go up?
Most teams pay less, because instances bill per second and scale down when traffic is low. The pricing page lists every rate.
04Can I keep my custom domains?
Yes. Add the domain to a service, update the DNS record, and the TLS certificate is issued automatically.