diff --git a/data/Dockerfiles/acme/Dockerfile b/data/Dockerfiles/acme/Dockerfile
index ce3fc171..944228ec 100644
--- a/data/Dockerfiles/acme/Dockerfile
+++ b/data/Dockerfiles/acme/Dockerfile
@@ -1,4 +1,4 @@
-FROM alpine:3.6
+FROM alpine:3.8
 
 LABEL maintainer "Andre Peters <andre.peters@servercow.de>"
 
diff --git a/data/Dockerfiles/phpfpm/Dockerfile b/data/Dockerfiles/phpfpm/Dockerfile
index 0ed81896..a23a767e 100644
--- a/data/Dockerfiles/phpfpm/Dockerfile
+++ b/data/Dockerfiles/phpfpm/Dockerfile
@@ -1,4 +1,4 @@
-FROM php:7.2-fpm-alpine3.7
+FROM php:7.2-fpm-alpine3.8
 LABEL maintainer "Andre Peters <andre.peters@servercow.de>"
 
 ENV APCU_PECL 5.1.12
diff --git a/data/Dockerfiles/watchdog/Dockerfile b/data/Dockerfiles/watchdog/Dockerfile
index 46130e5f..810b2af2 100644
--- a/data/Dockerfiles/watchdog/Dockerfile
+++ b/data/Dockerfiles/watchdog/Dockerfile
@@ -1,4 +1,4 @@
-FROM alpine:3.6
+FROM alpine:3.8
 LABEL maintainer "André Peters <andre.peters@servercow.de>"
 
 # Installation
diff --git a/data/Dockerfiles/watchdog/watchdog.sh b/data/Dockerfiles/watchdog/watchdog.sh
index ecfc6f07..b45efb85 100755
--- a/data/Dockerfiles/watchdog/watchdog.sh
+++ b/data/Dockerfiles/watchdog/watchdog.sh
@@ -206,7 +206,7 @@ postfix_checks() {
 clamd_checks() {
   err_count=0
   diff_c=0
-  THRESHOLD=6
+  THRESHOLD=5
   # Reduce error count by 2 after restarting an unhealthy container
   trap "[ ${err_count} -gt 1 ] && err_count=$(( ${err_count} - 2 ))" USR1
   while [ ${err_count} -lt ${THRESHOLD} ]; do
@@ -217,7 +217,8 @@ clamd_checks() {
     [ ${err_c_cur} -ne ${err_count} ] && diff_c=$(( ${err_c_cur} - ${err_count} ))
     progress "Clamd" ${THRESHOLD} $(( ${THRESHOLD} - ${err_count} )) ${diff_c}
     diff_c=0
-    sleep $(( ( RANDOM % 30 )  + 10 ))
+    # Don't check Clamd too often
+    sleep 1800
   done
   return 1
 }