Ethereum Fusaka Upgrade Event Analysis: Prysm Post-Event Report Reveals Root Cause of Failure

The Prysm development team recently released an after-action report detailing the mainnet anomaly event that occurred after the Fusaka upgrade on December 4, 2025. The issue once threatened the stability of the Ethereum network but was ultimately mitigated thanks to the diversity mechanism of clients.

The report shows that the problem occurred at epoch 411,392 (December 4, 21:49 UTC) after the Fusaka upgrade was activated. Prysm consensus clients triggered a large amount of repeated calculations of historical states when processing specific proof data, leading to rapid exhaustion of CPU and memory resources, causing nodes to experience denial-of-service (DoS)-type performance degradation. This was not a protocol design flaw but an implementation issue within the client under certain boundary conditions.

Affected Prysm validator nodes accounted for approximately 15% to 22.71% of the network. During the event, overall validator participation rate plummeted from normally over 95% to about 75%, resulting in the network missing 41 consecutive epochs, leading to a loss of about 382 ETH in attestation rewards and approaching finality loss. Prysm core developer Terence Tsao pointed out that the computational load of replaying historical states was extremely high, and multi-threaded parallel triggering significantly slowed node performance.

It is worth noting that the Fusaka upgrade itself was successful. The upgrade introduced PeerDAS (Peer Data Availability Sampling) technology, aiming to increase Layer 2 blob capacity by eight times. The upgrade process involved no downtime or consensus forks.

The reason the Ethereum network avoided more severe consequences was due to client diversity. Besides Prysm, ten other consensus clients such as Lighthouse, Teku, and Nimbus remained functioning normally throughout the process, allowing approximately 75% to 85% of validators to stay online and ensuring the network’s finality was not compromised. If a similar issue had occurred with a higher proportion of clients, the consequences could have been more severe, including suspension of Layer 2 aggregations and validator withdrawal blocks.

After the incident, the Ethereum Foundation quickly issued emergency guidelines. The Prysm team deployed temporary runtime fixes first and released permanent solutions in versions v7.0.1 and v7.1.0. By December 5, the network participation rate had recovered to nearly 99%, and the Ethereum mainnet fully resumed normal operation within 24 hours.