What Happens If the Base Sequencer Goes Down?
A Base sequencer outage temporarily stops new transactions from being ordered and confirmed, and there have been short incidents in the network's history. It is an availability problem, not a fund-safety one: assets remain controlled by your keys, users can force transactions through the Ethereum-side contracts, and because Base is an L2BEAT Stage 1 rollup, funds are exit-able even if the operator never comes back.
- ✓An outage stalls new transactions but cannot move or seize your funds
- ✓Transactions can be forced through the Ethereum L1 contracts as a fallback path
- ✓Base is Stage 1 on L2BEAT — the walkaway/exit test is already passed
- ✓During an incident: wait it out, avoid duplicate sends, and watch status.base.org
What the sequencer actually does
The sequencer is the operator-run server that orders incoming transactions, batches them, and submits them to Ethereum — think of it as a dedicated fast lane rather than the road itself. Today Coinbase runs Base's sequencer, which is the network's most centralised component and the honest answer to the "is Base decentralised" question. If that server has a bad upgrade, an infrastructure failure, or gets overwhelmed, the fast lane closes and wallets show transactions stuck in limbo.
But the sequencer never held your funds. The assets live in smart contracts anchored to Ethereum, so the operator's machinery going quiet does not give anyone access to them.
The two fallback paths
First, users can bypass the sequencer by submitting transactions directly to the rollup's contracts on Ethereum mainnet — slower and more expensive while it lasts, but permissionless by design. Second, even in a permanent operator failure, L2BEAT's Stage 1 designation means a user can run the exit process from the L1 contracts independently; the chain passes what the industry calls the walkaway test. Not every rollup has that designation, which is why it matters.
In the more common case — a short incident like the brief outages Base and Arbitrum have both experienced — the practical advice is boring: do not resubmit transactions repeatedly (they queue by nonce, as explained in the stuck-transaction guide), monitor the official status page, and wait. Transactions resume in order when the sequencer recovers.