Solana Nearly Frozen After Routing Error Took 29% of Stake Offline

On August 12, the Solana blockchain continued to produce blocks while a routing mishap at its largest infrastructure provider, Teraswitch, prevented validators representing roughly 29% of the network’s stake from voting.

Cause of the outage

Teraswitch identified a stale default route on a Miami edge router as the trigger. Routine maintenance by the transit provider re‑activated the route, and a mis‑configured routing policy applied a “no‑export” instruction that blocked traffic to Europe and Asia‑Pacific regions. The faulty route propagated across Teraswitch’s network, cutting Internet and inter‑site connectivity for twelve data‑center sites that host Solana validator and RPC nodes.

Impact on network performance

Independent monitoring by Metrika confirmed that the chain remained live but suffered degraded performance:

• Skipped leader slots rose above 32%.
• Transaction throughput fell to under 300 TPS, compared with the usual 1,100‑1,300 TPS.
• Finality was delayed for about 30 minutes, with newly produced blocks taking longer to become irreversible.

Despite the disruption, Solana avoided a full production halt because the offline stake stayed 4‑5% below the 33% finality‑halt threshold.

Response and future measures

Teraswitch deployed hardening on its provider side the same day, preventing an invalid route from disabling healthy local edge routers. The September post‑mortem noted ongoing work on additional route monitoring and highlighted the need for greater infrastructure diversity and observability.

Broader implications

The incident underscores concentration risk, as many validators shared a single provider. Although Teraswitch’s share of hosted stake had been reduced below 30% before the outage, the event shows how a single routing error can affect a sizable portion of the network.