integration-race: bump timeout from 20 to 30 minutes

There were still a few jobs runs were some tests (most recently:
test/integration/scheduler_perf/misc) timed out. We could split that up a bit
more, but as integration testing with race detection isn't something that needs
to complete quickly it's simpler to raise the timeout.

To prevent accidental long job runs when this individual timeout gets
reached by a higher number of packages, the job timeout gets reduced
from 2h (the default) to 90m.

Author: pohly

config/jobs/kubernetes/sig-testing/integration.yaml

@@ -73,6 +73,8 @@ presubmits:
     always_run: false # Has known failures.
     optional: true
     decorate: true
+    decoration_config:
+      timeout: 90m
     skip_branches:
     - release-\d+.\d+ # per-release job
     annotations:
@@ -92,7 +94,7 @@ presubmits:
         - runner.sh
         env:
         - name: KUBE_TIMEOUT
-          value: "-timeout=20m"
+          value: "-timeout=30m"
         - name: KUBE_RACE
           value: "-race"
         args:
@@ -219,6 +221,8 @@ periodics:
   cluster: k8s-infra-prow-build
   name: ci-kubernetes-integration-race-master
   decorate: true
+  decoration_config:
+    timeout: 90m
   extra_refs:
   - org: kubernetes
     repo: kubernetes
@@ -245,7 +249,7 @@ periodics:
       - name: SHORT
         value: --short=false
       - name: KUBE_TIMEOUT
-        value: "-timeout=20m"
+        value: "-timeout=30m"
       - name: KUBE_RACE
         value: "-race"
       # docker-in-docker needs privileged mode