Hassan Maishera
TL;DR
-
Sui Foundation published its incident review on Sunday, tracing three mainnet halts on Thursday and Friday to two separate bugs tied to its v1.72 upgrade.
-
The foundation said an interim fix it deployed on Thursday carried a known halt risk, the team agreed to restore service quickly, and the network went down again on Friday.
Two Separate Technical Failures Behind Network Downtime
The Sui Foundation has confirmed that three outages, which disrupted the Sui network across Thursday and Friday, were caused by two distinct software bugs introduced in the v1.72 upgrade.
In a post-incident review published Sunday, the foundation clarified the timeline of events and resolved uncertainty over whether a single root cause was responsible.
Following last week’s outages related to the 1.72 release, the Sui Core Team has completed an investigation and incident review, detailing what happened and the steps taken by validators to restart the network.
— Sui (@SuiNetwork) May 31, 2026
It confirmed that the disruptions were unrelated in their final trigger, although they originated from the same upgrade cycle.
First Outage Linked to Gas Accounting Flaw in v1.72 Update
According to the report, the first outage began around 10 a.m. ET on Thursday and lasted until approximately 4:30 p.m. ET. A second outage followed early Friday morning and was resolved later that day.
Both incidents stemmed from a flaw in the way the network calculated gas charges after the introduction of a new “address balances” feature in v1.72. The bug allowed transactions to be canceled for insufficient funds while still deducting those funds, creating negative balances that disrupted validator reconciliation processes and ultimately halted the chain.
Patch Introduced Instability that Triggered the Second Halt
The foundation said its initial emergency fix, deployed on Thursday, included a known low-probability issue. While intended to restore network activity quickly, that temporary patch introduced a new risk that materialized on Friday, triggering the second outage.
This meant the mitigation itself contributed to the instability that followed, as validators encountered a variant of the original accounting issue.
A separate and previously undisclosed issue caused the third outage on Friday evening, which lasted until around 10:20 p.m. ET.
This failure occurred during a validator restart required to apply the second fix. The restart exposed a latent bug in how the network preserves randomness state across epochs.
At the start of each epoch, Sui validators initialize a random number generation process used by certain applications. The foundation said too few validators completed the setup correctly after the restart. While the system correctly disabled randomness, a bug prevented validators from recording that state change, leaving the network unable to close the epoch and causing a third freeze.
Sui Foundation Deploys Fixes, Funds Unaffected
The foundation said both the gas accounting bug and the randomness-state bug have now been resolved. It emphasized that no user funds were ever at risk and that no finalized transactions were reversed during recovery.
It also introduced a mechanism to forcibly close stalled epochs, which was used once during incident resolution.
“As of now, validators have fully addressed the known issues caused by both the original gas-charging bug and the randomness-state bug, and network activity has resumed,” the foundation stated.
The report noted that AI agents integrated into production systems helped accelerate diagnosis by analyzing validator logs and generating metrics across the incidents.
It also highlighted that the v1.72 upgrade—intended to improve usability through features like address balances and gasless stablecoin transfers—created new execution paths where the bugs surfaced.
At the time of writing, SUI was trading around $0.88, down about 3.4% over 24 hours.
The Sui Foundation said it is now focusing on improving fault containment mechanisms so future bugs would degrade transaction-level functionality rather than halt the entire network. It added that strengthening resilience and limiting systemic failure remain a priority going forward.
