At 19:47 UTC, Sui saw a concentrated burst of order maintenance rather than a failure storm. One sender, 0xcde6db…6c0e04, submitted 466 of the minute’s 1,202 eligible transactions—38.8% of the total. Across the 1,195 transactions that could support a semantic action label, cancellations led with 360, followed by limit-order placements with 314 and price updates with 206. Together, those three behaviors accounted for 880 of 1,195 rows, or 73.6%.
The minute was unusually successful by the frozen corpus’s standards: 9 of 1,202 eligible transactions failed, a 0.75% rate versus 8.36% across the full corpus. The system structurally explained 1,176 rows, but that does not mean they all succeeded. Seven rows stopped before an application action could be safely attributed, and six of those arrived together at checkpoint 300555333 from 0xcde6db…6c0e04. Each reported InsufficientFundsForWithdraw; the detail for 2QUphncb says a withdrawal request exceeded available funds, so no application action is asserted.
The other three failures were more specific. Two attempted reward distributions and stopped at pool::distribute_rewards; one entered liquidity-position maintenance and stopped at collect_reward_with_return_. Their descriptions explicitly say earlier application effects were rolled back. These failures establish attempted maintenance, not completed reward transfers or withdrawals.
What the minute moved
The movement record is striking even though it cannot be converted into prices, profit, or cross-asset value. Sender-boundary holding records for SUI show 38,663.996511011 SUI in and 42.758907265 SUI out across 485 transactions, for a net of 38,621.237603746 SUI. USDC shows 78,858.095897 in and 1,069.170710 out across 34 transactions, for a net of 77,788.925187 USDC. WAL moved in the opposite direction: 60,590.513579646 in and 61,514.582923272 out across 24 transactions, for a net of −924.069343626 WAL.
Gas accounting is separate: SUI gas records show 0.103365308 SUI in and 1.060842382 SUI out, net −0.957477074 SUI. These accounting categories must not be added together because the same gas charge can also appear in sender-boundary records.
Three transactions explain the texture
Cjfc39Rq shows delegated trader authorization followed by a DeepBook limit-order call. It submitted a resting SUI/USDC order sized at 7,269.9 SUI and priced at 0.74565 USDC per SUI. The detail calls it resting exposure, not an exchange, and records a sender-boundary change of −0.002577652 SUI.
DkhRWKgH used an owner proof before an order-cancellation call for the NS/SUI book. The recorded inputs identify 1,051.4 NS of resting size at 0.01529 SUI per NS. That is evidence about the cancellation workflow and the order’s resting parameters; it does not establish that the order traded.
EoRogqVW submitted two price observations, 0.747255065328892 USDC per SUI and 1.338498815744766640128 SUI per USDC, both with the same expiry. The guarded oracle calls ran, but the transaction detail withholds any claim that a stored price actually changed.
The minute therefore reads as a tightly repeated maintenance cycle: cancellations, resting placements, and price housekeeping, with a small synchronized cluster of underfunded withdrawal attempts. The repeated shapes and cadence are consistent with automated execution, but the evidence does not establish who controlled the addresses or whether the activity came from one operator.
For a closer look, start with DkhRWKgH for cancellation mechanics, Cjfc39Rq for delegated order placement, and 2QUphncb for the six-transaction funding failure cluster.