v1.13.0

What's Changed

Storage & Reliability

• Block storage now reclaims space: unreferenced chunks are garbage-collected and overlapping writes coalesced, so volume growth stays bounded
• A full storage backend surfaces ENOSPC to the guest instead of stalling writes
• S3 storage refuses new writes at a pool free-space watermark and signals nearfull backpressure before the pool is exhausted
• Compaction now fsyncs its index and persists the segment counter before dropping a segment, so a crash mid-compaction no longer strands objects
• Objects with binary keys survive a Raft snapshot restore intact
• Legacy JSON snapshots restore correctly on upgrade
• Volume tags are no longer lost while the volume is live

Networking

• ALB and NLB endpoints resolve by DNS name
• EKS in-cluster DNS resolves from the VPC
• DNS records reconcile across tenants, so stale entries no longer linger
• Hardened DNS forwarder behaviour under upstream failure

Compute

• Root volume sizing is honoured at launch across EC2, EKS and ECS
• Ubuntu image catalog entries refreshed and bootable again

Security

• ECR registry operations now enforce IAM authorization
• Secret material removed from block storage logs

Load Balancing (ELBv2)

• A failed load balancer launch rolls back its managed security group and reports the underlying cause

Observability

• DNS service emits OpenTelemetry traces and metrics

Full Changelog: v1.12.1...v1.13.0