qa/crontab: update frequency and priority for rados nightlies

Although main rados nightlies are scheduled to run weekly, they
are so low in the queue that they take weeks to run, and when
they finally do, most of the jobs die (likely due to waiting in the
queue so long).

This commit increases the priority of main rados runs since
these are most important for debugging new failures. To compliment this
increase in priority, this commit also decreases the frequency of squid
rados runs from twice a week to once a week, since once is more than
enough. It also decreases the prioirity of squid runs.

Signed-off-by: Laura Flores <lflores@ibm.com>

qa/crontab/teuthology-cronjobs

@@ -71,7 +71,7 @@ TEUTHOLOGY_SUITE_ARGS="--non-interactive --newest=100 --ceph-repo=https://git.ce
 ## see script in https://github.com/ceph/ceph/tree/main/qa/machine_types
 
 # rados is massive and difficult to bring down to less than 300 jobs, use one higher priority
-00 20 * * 0        $CW $SS 100000 --ceph main --suite      rados -p 951
+00 20 * * 0        $CW $SS 100000 --ceph main --suite      rados -p 101
 08 20 * * 1        $CW $SS     64 --ceph main --suite       orch -p 950
 16 20 * * 2        $CW $SS    128 --ceph main --suite        rbd -p 950
 24 20 1 * *        $CW $SS    512 --ceph main --suite         fs -p 700
@@ -86,7 +86,7 @@ TEUTHOLOGY_SUITE_ARGS="--non-interactive --newest=100 --ceph-repo=https://git.ce
 
 # rados is massive and difficult to bring down to less than 300 jobs, use one higher priority
 #                                                                 -p 94-
-00 21 * * 0,4      $CW $SS 100000 --ceph squid --suite      rados -p 101 --force-priority
+00 21 * * 0        $CW $SS 100000 --ceph squid --suite      rados -p 700 --force-priority
 08 21 * * 1,5      $CW $SS     64 --ceph squid --suite       orch -p 100 --force-priority
 16 21 * * 2,6      $CW $SS    128 --ceph squid --suite        rbd -p 100 --force-priority
 24 21 * * 3,0      $CW $SS    512 --ceph squid --suite         fs -p 100 --force-priority