Add query to see jobs that need runners scaled up

[ZainRizvi]

This will be used by the autoscaler lambdas to figure out which instances may not have properly propagated the scale-up command (which happens fairly regularly)

The query is based on the "queued_jobs" query, with a few changes:

• Queues across all jobs in the pytorch and pytorch-labs organizations, which our runners support
• Checks only a specific time window for queued jobs, looking for jobs that have been queued for long enough to warrant intervention, while not being so old that github cancels the job

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
torchci	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Nov 19, 2024 11:36pm

[clee2000]

I would appreciate if you could update this with a code pointer in the future for where you use this, otherwise there's a good chance I'm going to think it's not used anywhere since there are no references in test-infra