[HUD] Look for queued jobs in past month only (#6166)

Make results for queued_jobs_by_label more similar
* Change queued_jobs_by_label to only look at jobs in the past weekly
like the queued_jobs query
* Use job.created_at instead of started_at for queued_jobs_by_label. I
don't think there's a difference for jobs that are queued, only jobs
that are running or completed

This helps resolve a problem where a job showed up in one HUD table but
not in the other

torchci/clickhouse_queries/queued_jobs_by_label/query.sql

@@ -1,9 +1,11 @@
 --- This query is used by HUD metrics page to get the list of queued jobs grouped by their labels
 WITH possible_queued_jobs as (
     select id, run_id from default.workflow_job where status = 'queued'
+    AND created_at < (CURRENT_TIMESTAMP() - INTERVAL 5 MINUTE)
+    AND created_at > (CURRENT_TIMESTAMP() - INTERVAL 1 WEEK)
 ), queued_jobs AS (
     SELECT
-        DATE_DIFF('second', job.started_at, CURRENT_TIMESTAMP()) AS queue_s,
+        DATE_DIFF('second', job.created_at, CURRENT_TIMESTAMP()) AS queue_s,
         CONCAT(workflow.name, ' / ', job.name) AS name,
         job.html_url,
         IF(
@@ -27,7 +29,7 @@ WITH possible_queued_jobs as (
         and workflow.id in (select run_id from possible_queued_jobs)
         and workflow.repository. 'full_name' = 'pytorch/pytorch'
         AND job.status = 'queued'
-        AND job.started_at < (CURRENT_TIMESTAMP() - INTERVAL 5 MINUTE)
+        AND job.created_at < (CURRENT_TIMESTAMP() - INTERVAL 5 MINUTE)
         /* These two conditions are workarounds for GitHub's broken API. Sometimes */
         /* jobs get stuck in a permanently "queued" state but definitely ran. We can */
         /* detect this by looking at whether any steps executed (if there were, */