The Incus team has just announced the release of version 6.12 of its container & virtual machine manager, bundled with several long-requested quality-of-life improvements and a handful of performance boosts.
The main highlight – virtual machines can now dynamically increase their RAM allocations. Until today, Incus supported ballooning memory down, but not up—a mismatch with container behaviour that occasionally forced administrators to reboot guests just to satisfy peak workloads.
Version 6.12 introduces genuine memory hot-plug: simply raise the limits.memory value, and the guest sees the extra gigabytes immediately, no restart required.
Moreover, network security rules have grown easier to read and maintain. The new address-set abstraction lets operators group any mix of IPv4 and IPv6 addresses under a project-scoped label and then reference that label from access control lists.
Also on the network side, when a bridged forward sends external ports 80/443 to an internal service on different ports, outbound packets from the instance can now be SNAT-ed to match the public address and port.
Think of it as creating a named alias, such as cloudflare-dns, and re-using it across multiple ACL rules instead of sprinkling raw IPs everywhere.
Incus logging used to be a binary choice—local syslog or a single Loki endpoint. Version 6.12 replaces that limitation with a plug-in architecture that supports multiple targets, each of which can be Loki or classic syslog.
Administrators can now fine-tune which event classes (such as lifecycle, ACL hits, and generic log lines) flow to which destination and at what severity. In larger deployments, this means, for instance, shipping high-volume debug traffic to a Loki cluster while reserving syslog for warnings and above.
On the command-line side, the Incus client gains server-side filtering for instances, images, custom volumes, and profiles. Instead of pulling giant lists over the wire and filtering them locally, the CLI offloads the work to the daemon, resulting in less network traffic and less SQLite churn on busy hosts.
Lastly, the Incus team calls out performance improvements for installations with many snapshots and for ZFS users, though the post does not publish raw numbers. Still, folks running snapshot-heavy dev environments should notice snappier start-up and list operations.
For more information about the Incus 6.12 container and virtual machine manager changes, visit the release announcement or check out the full changelog.
Users are encouraged to try out these new features by visiting the Incus online platform, which provides a hands-on experience with the latest version.
