Retention
Retention rules define how long raw and computed telemetry remains available for live access, historical queries, operational review, and Analytics use cases.
Choose retention based on compliance, cost, customer commitments, and product requirements.
Raw And Computed Retention
Raw and computed streams may need different retention choices:
| Stream | Retention consideration |
|---|---|
| Raw telemetry | Useful for audit, troubleshooting, replay, and validating computed signals. It may be high volume. |
| Computed streams | Useful for applications, customer views, operational signals, and Analytics. They may be smaller and kept longer. |
Keeping raw data forever is rarely the right default for high-volume IoT fleets. A common approach is to keep raw telemetry for the period where debugging and replay matter most, then keep computed streams for longer-term operational access.
What To Consider
When defining retention, consider:
- how long customers expect historical data to remain available,
- whether compliance requires deletion or preservation,
- how often teams troubleshoot from raw telemetry,
- which computed signals are needed for long-term operations,
- where data is hosted and who controls it,
- the cost of storing high-volume timestamped records.
Hosting And Data Control
Infuse DB supports cloud-managed, private, and on-prem hosting choices. Retention expectations should be defined alongside the hosting model because data control, operational responsibility, and compliance requirements vary by deployment.