bump results watcher watcher container to 3Gi (#3093)

Over the course of about a week we have seen the results watcher on prd-rh01  have to OOM about once a day.  During the day, we have observed garbage collection, and seen mem usage go up and down some.  But as prd-rh01 is our most used cluster, it is the one that has bumped into the 2Gi limit too often.  The other 3 clusters, while still having activity, have not observed leak behavior.  We have engaged upstream with some profiling data from our personal clusters to see if there is some sort of edge case.  But the simpler step first step to bump prod, and then continue to monitor, and see if we have just hit a new limit with more onboarding.

components/pipeline-service/production/base/bump-results-watcher-mem.yaml

@@ -0,0 +1,4 @@
+---
+- op: replace
+  path: /spec/template/spec/containers/1/resources/limits/memory
+  value: "3Gi"

components/pipeline-service/production/base/kustomization.yaml

@@ -25,6 +25,11 @@ patches:
       kind: Deployment
       name: pipeline-metrics-exporter
       namespace: openshift-pipelines
+  - path: bump-results-watcher-mem.yaml
+    target:
+      kind: Deployment
+      namespace: tekton-results
+      name: tekton-results-watcher
   - path: update-tekton-config-pac.yaml
     target:
       kind: TektonConfig

components/pipeline-service/production/stone-prd-m01/deploy.yaml

@@ -1514,7 +1514,7 @@ spec:
         resources:
           limits:
             cpu: 250m
-            memory: 2Gi
+            memory: 3Gi
           requests:
             cpu: 100m
             memory: 64Mi

components/pipeline-service/production/stone-prd-rh01/deploy.yaml

@@ -1514,7 +1514,7 @@ spec:
         resources:
           limits:
             cpu: 250m
-            memory: 2Gi
+            memory: 3Gi
           requests:
             cpu: 100m
             memory: 64Mi