Base recovers from a 33-minute mainnet block-production halt

What happened

Base's active sequencer fell behind, and an automated Conductor handoff selected a sequencer that was still being provisioned and could not produce blocks. Mainnet block production stopped for 33 minutes.

What changed

The incident established a realized sequencer failover and provisioning failure capable of interrupting Base mainnet block production, deposits, and withdrawals. Base said it would change cluster admission behavior and testing.

What did not change

Base reported full network recovery by 06:40 UTC. The postmortem did not report a bridge-contract exploit, asset loss, or chain reorganization.

Confirmed

• Base mainnet block production stopped at approximately 06:07 UTC on August 5, 2025.
• The status incident identified deposits, withdrawals, block production, and Flashblocks as affected components.
• The team manually paused Conductor and transferred leadership to a healthy sequencer.
• Base reported full recovery after 33 minutes.

Still open

• The status postmortem states planned infrastructure and testing improvements but does not prove when each corrective action was deployed.
• The receipt does not quantify transaction-level deposit or withdrawal delays after block production resumed.

What it means for an advisor

• Review the coinbase-bridge memo's availability and withdrawal assumptions against a realized sequencer failover outage.
• Record that canonical bridge access can become temporarily unavailable during a Base sequencing halt even without a bridge-contract exploit.
• Verify whether current operational controls now prevent an unready sequencer from becoming leader.