Investigate server failures during 8.4 tests

Author: petyaslavova

.github/actions/run-tests/action.yml

@@ -142,6 +142,7 @@ runs:
         sudo apt-get install -y redis-tools
         echo "Docker Containers:"
         docker ps
+        echo "Cluster nodes:"
         redis-cli -p 16379 CLUSTER NODES
       shell: bash
 

dockers/sentinel.conf

@@ -1,5 +1,8 @@
 sentinel resolve-hostnames yes
 sentinel monitor redis-py-test redis 6379 2
-sentinel down-after-milliseconds redis-py-test 5000
-sentinel failover-timeout redis-py-test 60000
+# Be much more tolerant to transient stalls (index builds, GC, I/O)
+sentinel down-after-milliseconds redis-py-test 60000
+# Avoid rapid repeated failover attempts
+sentinel failover-timeout redis-py-test 180000
+# Keep it conservative: sync one replica at a time
 sentinel parallel-syncs redis-py-test 1