This one took about 31.9× the usual for KMDCL/LCC. Waiting longer than usual is not in itself a failure: a swap waits on confirmations from both chains, and a slow block or a busy network stretches it.
Nothing was lost. The payment was never claimed: it was spent without the preimage, after the timelock, which is the refund path back to the payer, in 9020925c5861c3320ae623bc29b2dca8c7883c4dcb3f34dd746abab27cf0395c.
The swap stopped at TakerPaymentWaitForSpendFailed.
utxo_common:3390] Timeout { wait_until: 1789951362, now: 1789951364 }
Recorded by the swap's own protocol log, which is the only source that knows: the aggregator reports failed and completed swaps identically. Nothing changed hands, so it sets no reference price.
Only completed swaps are checked against a chain, because the check asks whether the agreed trade actually settled. This one did not complete, so there is nothing of that kind to confirm.
Reported by kdf-statsdb.