The satellite cannot cancel the upload - it’s not proxying the traffic. The customers’ uplink uploads directly to your node. So it’s possible, that some customer didn’t cancel uploads, but the connection will be terminated anyway (up to 48h as far as I know).
So far I have several ideas:
- for some reason the node could report to the satellite as full;
- the public IP has changed and DDNS didn’t update it in time, but in this case the node will be rendered as offline;
- there was a temporary connection issue and the node was removed from the hot cache on the satellite and suggested to the customers not so often as before, until it again will not be added to the hot cache;
Could you please check - did it have audits requests between these two dates?
Also please check for missing audits: