From 41d11b4bbe114137181c4e7f38120b4cf3ee1fe4 Mon Sep 17 00:00:00 2001 From: "cortex-ai-agents[bot]" <279748396+cortex-ai-agents[bot]@users.noreply.github.com> Date: Mon, 6 Jul 2026 07:37:15 +0000 Subject: [PATCH] docs(apis): update FlavorGroupCapacity status fields after #983 Co-Authored-By: Claude Opus 4.7 --- docs/apis.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/apis.md b/docs/apis.md index 6a1c2bdf7..b1dd1367d 100644 --- a/docs/apis.md +++ b/docs/apis.md @@ -142,7 +142,7 @@ For more details on how committed resources interact with reservations, see [com kubectl get flavorgroupcapacities ``` -FlavorGroupCapacity caches pre-computed capacity data for one flavor group in one availability zone. One CRD exists per (flavor group × AZ) pair, maintained by the capacity controller on a fixed interval. The spec identifies the flavor group and AZ; the status holds per-flavor slot counts (`PlaceableVMs`, `PlaceableHosts`, `TotalCapacityVMSlots`, `TotalCapacityHosts`), aggregate fields (`CommittedCapacity`, `TotalCapacity`, `TotalInstances`), and a `LastReconcileAt` timestamp. The capacity API reads these CRDs instead of probing the scheduler on each request. +FlavorGroupCapacity caches pre-computed capacity data for one flavor group in one availability zone. One CRD exists per (flavor group × AZ) pair, maintained by the capacity controller on a fixed interval. The spec identifies the flavor group and AZ; the status holds per-flavor slot counts (`PlaceableVMs`, `PlaceableHosts`, `TotalCapacityVMSlots`, `TotalCapacityHosts`), aggregate fields (`CommittedCapacity`, `CommittedCapacityBytes`, `TotalCapacity`, `FreeCapacity`, `ExclusivelyFreeCapacity`, `ExclusivelyFreeSlots`, `RunningInstances`, `RunningResources`, `SmallestFlavorName`), and a `LastReconcileAt` timestamp. Where flavor groups share hosts, remaining capacity is fairly split across overlapping groups using a round-robin algorithm so that no group's exclusively-free capacity exceeds the actual installed resources. The capacity API reads these CRDs instead of probing the scheduler on each request. ### ProjectQuota