What is Scale to Zero?
Scale to Zero reduces a service's running instance count to zero when there is no load. A trigger mechanism is needed to start it again for new requests. This is a resource operating model, not deletion of your application or database.
Benefits and tradeoffs
| Concern | What it means |
|---|---|
| Idle resources | Running compute can decrease during periods without traffic; savings depend on the service's billing model. |
| Cold start | The first request after inactivity may wait for startup. Measure the actual behavior with your code and data. |
| Data | Do not rely on instance memory for persistent data. Use a database/durable storage with a backup plan. |
| Continuous work | Workers, long-lived connections and background tasks need suitable triggers; do not assume they run while there are zero instances. |
| Remaining costs | Storage, databases, network addresses or other charges may remain. Zero instances does not mean a zero bill. |
Applying the concept to DLEZ
This is a concept guide. Availability for specific DLEZ products/plans has not been confirmed here, so this guide does not publish activation steps. Do not design production around an assumption that an app automatically sleeps or wakes.
VPS Stop/Start and app redeployment are separate actions; they do not establish Scale to Zero support. Use the existing resource configuration and Billing guides to manage current usage.
If your service supports it, what should you verify?
- Confirm product/plan availability, idle conditions and wake-up triggers, including both public and private calls.
- Measure the first request after inactivity, handle timeouts/retries appropriately and verify background work.
- Check data after sleeping and waking; do not use temporary memory as your primary data store.
- Review actual costs and confirm charges that remain with zero compute instances.