Post #238

The Arbitrum bridge was locked earlier today due to an automatic monitoring process run by L1 validators. The locking process relies on many major Arbitrum RPCs and ensures that the L1 and Arbitrum state are synced. The locking mechanism is meant as a last resort to prevent exploits. In this case, the locking was a false positive. An extremely rare edge case occurred where multiple RPCs served inconsistent data between transactions and state. This is not an error by the RPC providers because there is no guarantee of temporal consistency between endpoints, especially on a high throughput network like Arbitrum. However, it is extremely rare that many RPCs agree exactly on their out-of-date data, which caused a locking condition to be hit. The overall locking mechanism is conservative because losing user funds is a far worse outcome than occasional downtime. This was the first locking incident in over one year of continuous operation. Thank you to the Arbitrum team for prompt assistance in looking into the issue. Improvements are being made to the monitoring process to prevent the same false positive from triggering in the future. All user funds are safe.