fix(nvd-mirror): increase deadline to 2h (google#2538)

Given the cron job runs every 2 hours, there's nothing gained by
arbitrarily terminating it before that. At 260K records over 130 pages,
the best case scenario for *retrieval* alone is ~780 seconds (and
growing) and with intermittent failures and retries, could exceed an
hour.

And that's before the time taken to split by year.

deployment/clouddeploy/gke-workers/base/nvd-mirror.yaml

@@ -8,7 +8,7 @@ spec:
   failedJobsHistoryLimit: 3
   jobTemplate:
     spec:
-      activeDeadlineSeconds: 2100
+      activeDeadlineSeconds: 7200
       template:
         spec:
           tolerations: