Chain halt (network outage)

Halts happen when the consensus process cannot move forward. Typical causes are a software bug triggered by an unusual transaction, a flood of spam that overwhelms validators, a disagreement between client versions or not enough stake online to reach the required majority. On many proof-of-stake chains, recovery requires validators to agree on a patched release and restart from the last good block, coordinated through public channels.

Outages matter because everything built on the chain freezes with it: exchanges suspend deposits and withdrawals, DeFi positions cannot be adjusted and price oracles go stale, which can trigger unfair liquidations once blocks resume. Solana suffered several multi-hour outages in its early years, and halts on Cosmos-based chains and some layer 2 sequencers have also made headlines.

Chains make different trade-offs: some prefer to stop rather than risk conflicting histories, while others keep producing blocks but delay finality. A halt does not usually mean funds are lost, but a restart decided by a small group of validators shows how much coordination a network actually depends on. Running several independent client implementations reduces the chance that one bug stops everything.