Limits & quotas
Every platform boundary on one page. Where a limit is a default rather than a hard ceiling, it says so; project quotas in particular are raised on request.
Deployment size
Set per app in Settings; every deployment of the app runs at this size. See CPU, memory & storage.
| Resource | Values |
|---|---|
| CPU | 0.5, 1, 2, or 4 cores |
| Memory | Steps from 0.5 to 16 GB, between 1× and 4× the CPU cores (see the matrix) |
| Storage | 5–100 GB, in 5 GB steps |
| Default size (new apps) | 0.5 CPU · 1 GB memory · 5 GB storage |
Storage is all-inclusive (volumes, database data, and backup history share the one budget) and can grow but never shrink. Size changes apply on the next deploy of a stopped deployment (how to resize).
Project quotas
Counted across a project's running deployments; enforced when a deploy is submitted. Defaults shown; contact support to raise any of them.
| Quota | Default |
|---|---|
| Total CPU | 2 cores |
| Total memory | 4 GB |
| Total storage | 25 GB |
| Running deployments | 5 |
Stopping a deployment releases its CPU, memory, and deployment slot. Its storage stays counted until the deployment is deleted.
Manifest
| Rule | Limit |
|---|---|
| Services per app | 1–32 |
| Volumes per app | 32 |
| Ports | 16 per service, 64 per app; each unique within its service and in the range 1–65535 |
| Environment variables | 128 per service |
| Docker build arguments | 64 per service |
| Volume mounts | 32 per service |
| Service dependencies | 32 per service |
| Healthcheck command items | 16 |
| Managed database services per deployment | 4 |
| PostgreSQL versions | "17" |
| Service and volume names | 1–64 characters: letters, digits, _, -; must start with a letter or digit (naming rule) |
Public-path suggestions (auth.public_paths) | 10 per manifest (suggestions only) |
Access
Set per app in Settings; see Access control.
| Rule | Limit |
|---|---|
| Public paths per app | 10 |
| Public path length | 128 characters |
| Public path form | An exact path (/healthz) or a prefix ending in /* (/api/*) |
Healthchecks
Full field reference in Container services: healthchecks.
| Field | Range | Default |
|---|---|---|
interval | 1s–5m | 10s |
timeout | 1s–5m | 3s |
retries | 1–20 | 3 |
start_period | 0s–5m | 0s |
Logs
| Log | Behaviour |
|---|---|
| Build logs | Captured for every build, success or failure, and kept as the deployment's build history. Very long build output keeps its final portion: the end of the log, where errors live. |
| Container logs | Fetched on demand from the running containers, newest first, with paging back through history. |
| Crash output | When a deployment fails, the final output of its containers is captured and kept with the deployment. |
Logs belong to their deployment and are removed when the deployment is deleted.
Where to go next
- CPU, memory & storage: choosing and changing sizes.
- Manifest reference: every valid field.