The Big Picture

This was the quietest week of the cycle — the holiday lull, spanning New Year's, with just 12 threads and 19 messages on openstack-discuss. There were no releases, no CVEs, and no major decisions. The week was dominated by operator support questions (production-cluster validation, a stuck Watcher audit) and a fresh open-source community tool, with a couple of substantive technical seeds planted by Sean Mooney and a notable openstacksdk memory-leak report carrying over into the new year. Treat this as a breather issue.

Development & Technical Decisions

openstacksdk Prometheus metrics memory leak. Sam Clippinger's detailed report (also active the following week) is the most technically important item: the openstacksdk records per-request Prometheus counters and histograms keyed on the full request URL, with no way to disable it short of uninstalling prometheus_client — which is pulled in transitively via oslo.metrics and oslo.messaging. Because nova-scheduler queries Placement with each VM's UUID, every unique URL gets its own counter, and at high VM churn this drives steady memory growth (his Caracal build farm sees nova-scheduler workers reach ~8 GB and get OOM-killed). He asks whether keystoneauth1's Adapter/CloudRegion should support a collect-timing-style off switch. Operators running at scale should be aware.

Nova ideas from Sean Mooney. Mooney seeded several Nova discussions over the break — using $aggregates with JsonFilter, unified-limits combined with PCI alias support (clarifying that resources:CUSTOM_NVIDIA_T4 should not be hand-added to flavors), and a cross-project heartbeat improvement to check whether the main thread actually responds to a RabbitMQ ping (touching Nova, Cinder, and Neutron). Several of these had no readable bodies in the archive but signal active design thinking that resurfaces in later weeks. Karol Klimaszewski also opened a Nova "ephemeral storage potential features" thread.

Heads Up / Action Needed

• Puppet modules losing a core. Alan Bishop announced his resignation from the puppet-cinder, puppet-glance, and puppet-manila core teams due to a change in work status. That's reduced review capacity across three Puppet modules — interested contributors should step up.
• Neutron bug deputy report (week of 2025-12-15) from Bence Romsics flagged two High-severity items: extremely slow functional tests using SqlFixture (fix in progress) and the highest-priority chassis gateway port failing to bind after a chassis restart/crash (WIP fix), plus several Medium and Incomplete bugs.
• Mind the noise. A "SECURITY BUG" thread and a "Bug reports" thread were bounty-seeking / impatient follow-ups from "Umbrella Corporation" with no actionable content — not real advisories.

Community & Events

The week's bright spot was a community contribution: SnapSentry, an MIT-licensed, Go-based open-source tool from Aravindh Murugesan to automate Cinder snapshot lifecycle management (scheduling and rotation, which Cinder leaves to the operator). It's metadata-driven (tag volumes, no central DB), does "atomic-ish" parallel snapshots of all volumes attached to a VM, uses hybrid concurrency to avoid API throttling, cleans up orphaned "zombie" snapshots, and supports restricted application credentials for least-privilege operation. The author positioned it as a lightweight alternative to commercial tools like Trilio and welcomed feedback.

On the support side, Tarun asked for help validating a production cluster while migrating from VMware to OpenStack 2025.2 — a reminder of how much VMware-migration traffic continues to flow into the lists. Mohd Atif Ansari reported a Watcher continuous audit stuck in PENDING with no action plans generated despite all services running, a recurring Watcher pain point that drew several replies from the community. A Nova "ephemeral storage potential features" thread (Karol Klimaszewski) opened the door to a discussion that would continue into January. Most project teams remained on their holiday meeting cancellations through the start of the new year, and the TC's weekly IRC meeting stayed paused — so expect the cadence, and the volume, to pick back up sharply in the following issue once contributors return from the break.