Base said it experienced two mainnet block production interruptions on June 25 and 26, lasting about 116 minutes and 20 minutes, respectively. According to Foresight News, the network said the incidents did not affect on-chain asset security and that user funds remained available.
Base attributed the disruption to a defect in the sequencer’s block-building logic. After a transaction execution failure, the system did not properly clear historical journal state, causing gas calculation errors when subsequent valid transactions executed. This led to invalid state-transition blocks and halted block production across the Layer 2 network.
During the outages, Base reported a complete stop in block production, transactions failing to be included on-chain, and mempool congestion. User-submitted eth_sendRawTransaction requests continued to return errors.
Base said it restored block production after applying a patch (PR #3806). However, it noted that an engine reset race condition during the sequencer cluster restart hindered recovery synchronization and indirectly contributed to the brief outage the following day.
Looking ahead, Base said it will strengthen protocol-level fuzz testing and stress testing to identify abnormal transaction paths earlier, upgrade monitoring and operations systems, and introduce a graceful recovery mechanism to improve restoration speed in similar incidents.