Woofun AI reports that a sequencer bug triggered two consecutive outages on the Coinbase layer-2 network Base last week, halting block production for a combined 136 minutes. The engineering team identified a flaw in block-building logic where "stale journal state" persisted following a transaction validation failure, preventing the network from clearing accessed accounts and storage slots. Base operates a single sequencer, a centralized component that dictates transaction order and has historically caused disruptions on other chains like Arbitrum, OP Mainnet, and zkSync Era.

The first incident on Thursday lasted 116 minutes, while a second outage on Friday persisted for 20 minutes, creating a complete stop in new layer-2 blocks. Validator nodes could not progress past the invalid block until sequencing was restored, as the system erroneously retained journal data despite the transaction failing execution as expected.

Woofun AI data shows the network secured just under $11 billion, making it the second-largest layer-2 by total value secured according to L2beat.

Engineers resolved the immediate crisis by patching sequencers to ensure journal states updated correctly during execution, though mitigation took longer than anticipated due to unrelated infrastructure conditions. A subsequent "race condition" after the system reset prevented sequencers from catching up, directly causing the second outage. This marks the third sequencer-related halt for Base, following a 17-minute stoppage in September 2024 and a roughly 30-minute disruption in August 2025.

To prevent recurrence, the team plans to enhance protocol "fuzz testing" by bombarding the system with random, malformed, or unexpected inputs to expose hidden bugs. They also intend to build "graceful recovery" mechanisms so validator nodes do not require manual restarts during future incidents. This pattern of centralized failure underscores the inherent fragility of single-node sequencing architectures in high-value environments.