Overview of Two Outages
Base mainnet experienced two block production interruptions on June 25 and 26, 2026, lasting approximately 116 minutes and 20 minutes respectively. The official postmortem report confirmed that user assets remained secure and on-chain funds were always accessible. During the outages, the network suffered complete block production halt, inability to include transactions, and mempool congestion. Users calling eth_sendRawTransaction received persistent errors.
Root Cause: Sequencer State Management Bug
The core issue resided in the sequencer's block construction logic. After a failed transaction execution, the system failed to properly clear the historical journal state, causing subsequent valid transactions to compute gas incorrectly and generating invalid state transition blocks. These invalid blocks effectively stalled the entire L2 network. The official patch (PR #3806) successfully fixed the issue and resumed block production. However, a race condition during sequencer cluster restart hindered synchronization recovery, which indirectly contributed to the second outage the following day.
Future Improvements and Recovery Strategy
The Base team announced plans to strengthen protocol-level fuzz testing and stress testing to detect abnormal transaction paths earlier. They will also upgrade monitoring and operational systems and introduce a more graceful recovery mechanism to enhance the network's ability to recover quickly from similar failures, minimizing impact on user transaction experience.