Completed
Maintenance ended
Updated
The maintenance is complete. All clusters are now running the patched kernel, and CVE-2026-98432 is fully fixed across our infrastructure. The temporary mitigations applied earlier have been replaced by the permanent fix.
Applications restarted automatically as their clusters came back online. If anything isn't behaving as expected, restarting it from the dashboard usually resolves it. If not, our support team is here to help.
Thank you for your patience while we kept the platform secure.
Updated
Rigel, Violet & Ponyo cluster updated and released.
Updated
Rigel, Violet & Ponyo cluster updating. Estimated time: 4 minutes and 29 seconds.
Updated
Suki & Iara cluster updated and released.
Updated
Suki & Iara cluster updating. Estimated time: 4 minutes and 53 seconds.
Updated
Nasa & Tsukasa cluster updated and released.
Updated
Nasa & Tsukasa cluster updating. Estimated time: 4 minutes and 27 seconds.
Updated
Vanda & Argon cluster updated and released.
Updated
Vanda & Argon cluster updating. Estimated time: 4 minutes and 32 seconds.
Updated
Cosmo & Nebula cluster updated and released.
Updated
Cosmo & Nebula cluster updating. Estimated time: 3 minutes and 40 seconds.
Updated
Orion cluster updated and released.
Updated
Orion cluster updating. Estimated time: 3 minutes.
Updated
Aurora cluster released; we took 3 minutes and 47 seconds longer than expected, our sincere apologies.
Started
A critical vulnerability was recently disclosed in the Linux kernel's TCP/IPv4 networking stack (CVE-2026-98432, CVSS 9.8). Under certain conditions, it could allow remote code execution on the host without any credentials.
What we've already done
Mitigation patches were applied across our entire infrastructure as soon as the vulnerability was disclosed. They significantly reduce the risk, but they are a temporary safeguard and do not fix the flaw itself.
Why we're updating now
The only complete fix is a patched kernel from our OS vendor (Canonical/Ubuntu). Rather than deploying it the moment it was published, we waited until it had been validated on more than 2,000 enterprise servers running Ubuntu. Now that it has proven stable, we're rolling it out across our platform. A new kernel only takes effect after a reboot, which is why this maintenance is needed.
What to expect
- How: a rolling reboot in small batches of clusters, so the platform is never down all at once.
- Impact: each application may be unavailable for a few minutes while its cluster restarts (up to ~15-20 minutes in the worst case). Applications come back up automatically.
- Action required: none. If your application keeps state in memory, make sure it handles restarts cleanly.
We'll post updates here throughout the maintenance. If you have any questions, our support team is here to help.
Thank you for your understanding.
Resolved