Add Disk Usage to the VM ranking panel

[traviswu-bigstack]

What type of PR is this?

/kind feature

What this PR does / why we need it

Adds a fourth option to the VM ranking panel on /home/chart: Disk Usage, ranking VMs by bytes allocated in the storage pool. The card previously offered CPU, Memory and disk I/O, but nothing about how much space a VM actually occupies — the top-line ask from the customer request behind bigstack-oss/cubecos#1449.

Labelled "Disk Usage" rather than "Storage Usage" because it sits beside "Cpu Usage" and "Memory Usage", where the user is picking a resource, not an API metric name.

Which issue(s) this PR fixes

Part of bigstack-oss/cubecos#1449. Needs bigstack-oss/cube-cos-api#663 (serves storageUsage) and bigstack-oss/cube-cos-openapi#115 (the enum).

Special notes for your reviewer

The SDK diff is generated, not hand-written — pnpm api:generate against the bumped spec submodule. Worth knowing that the metric type enum appears twice in the generated client, GetMetricByHostOrVmMetricTypeEnum and GetMetricByTypesMetricTypeEnum. Editing only one of them still type-checks, which is a quiet way to leave the other caller unable to name the value; regenerating keeps both in step.

Verified on a live single-node cluster: the dropdown offers Disk Usage and the panel renders eight VMs with sparklines. Note this panel was previously blank for every metric on that cluster — cube-cos-api was querying the retired monasca database — so it is worth re-checking against a cluster that has bigstack-oss/cube-cos-api#663.

Additional documentation

[traviswu-bigstack]

Build / Check are red for a pre-existing reason, not this change

Both fail on 12 TS2741: Property 'deviceProfile' is missing errors in src/mocks/gpu.ts — no file this PR touches.

The cause is that this repo's spec submodule was pinned at a34418d, several commits behind the spec's own develop. Picking up storageUsage means moving the pin forward, which also picks up feat(gpu): document a pgpu card's device profile — and the GPU mocks here were never updated for it.

I verified this is drift rather than anything in this PR: bumping the submodule to the spec's develop tip without my commit and regenerating produces the same 12 errors. develop here has no deviceProfile in those mocks at all.

So it needs a separate fix — adding deviceProfile to the twelve GPU mock objects — which belongs with the GPU feature, not inside a storage PR. Happy to file it as its own issue if that's useful. Until then this PR stays draft; it already depends on bigstack-oss/cube-cos-api#663 and bigstack-oss/cube-cos-openapi#115 landing first.