Run respdiff also for scheduled and web pipelines

respdiff needs to be run regularly to identify problems with query
responses discrepancies sooner than after tagging a release.

Contrary to the main branch, which hosts the BIND 9 Development Version
and thus sets MAX_DISAGREEMENTS_PERCENTAGE variable to 0.5, branches
hosting BIND 9 Stable Versions have it set to 0.1, which provides only
tiny room for non-timeout response disagreements between the baseline
version and version under test.

(cherry picked from commit 561b58196b)

.gitlab-ci.yml

@@ -1377,11 +1377,13 @@ coverity:
 
 respdiff:
   <<: *base_image
+  <<: *schedules_tags_web_triggering_rules
   stage: system
   variables:
     CC: gcc
     CFLAGS: "${CFLAGS_COMMON} -Og"
     BIND_BASELINE_VERSION: v9_11_24
+    MAX_DISAGREEMENTS_PERCENTAGE: "0.1"
   script:
     - ./configure --without-make-clean
     - make -j${BUILD_PARALLEL_JOBS:-1} V=1
@@ -1396,14 +1398,13 @@ respdiff:
   needs:
     - job: tarball-create
       artifacts: true
-  only:
-    - tags
   artifacts:
     paths:
       - refbind
+      - bind-qa/bind9/respdiff
     untracked: true
     expire_in: "1 day"
-    when: on_failure
+    when: always
 
 # "Stress" tests