LKE-E outage: subnet IDs in scientific notation broke new clusters

Starting October 17, 2026, new Linode Kubernetes Engine-Enterprise clusters failed to provision after VPC subnet IDs crossed 1,000,000: the Cloud Controller Manager parsed 7-digit IDs as floats in scientific notation. Akamai enforced strict integer formatting and patched resources, mitigating the incident at 19:49 UTC on October 20, 2026.

• New LKE-E clusters hung while existing clusters and scaling kept working
• Cause: subnet IDs parsed as floats and rendered as 1e+06
• Fix enforcing strict integer formatting shipped October 20, 2026
• Postmortem update published October 21 at 15:49 UTC