RK3576 · 72-hour dual-CV long-run observation
8 channels · 16 business-task bindings · 5 FPS/task
Observation result
| Configured channels | Task bindings | Duration | Target FPS/task | Minimum / average / maximum FPS | Result |
|---|---|---|---|---|---|
| 8 | 16 | 72 h | 5 | 5 / 5.098 / 5.17 | PASS |
Scope: Results apply to the listed channel counts and controlled local-loop input; maximum capacity, RTSP resilience, and restart recovery were not measured in this run. See the multi-platform evidence notes and methodology for complete execution conditions.
Sampling continuity and integrity
| Observed / expected samples | Coverage | First-sample lag | Final-sample lag | Maximum gap | Collector errors | Incomplete bindings | Missing binding telemetry | Open critical incidents | Integrity |
|---|---|---|---|---|---|---|---|---|---|
| 4316 / 4320 | 99.91% | 111.99 s | 59.867 s | 60.058 s | 0 | 0 | 0 | 0 | PASS (8/8) |
Workload and resources
| Maximum discard rate | Peak CPU | Peak memory | Disk first → last | Disk minimum / peak | Disk changes |
|---|---|---|---|---|---|
| 0% | 46% | 30% | 14% → 15% | 14% / 15% | 5 |
Provenance
Source: 44209759f450e96cda265acfa8bc6d17a1138888 · tree 5cbdefeaefaf642407356c22c271ccc7d57935b0
Multi-platform report and evidence notes · canonical data · methodology