service/cluster: monitor health of agents

When an agent is unhealthy only notify about that,
and suppress notifications about other services on that agent
not running.

Signed-off-by: Tobias Wiese <tobias@tobiaswiese.com>

1 files changed, 20 insertions, 0 deletions

diff --git a/zones.d/master/cluster-health.conf b/zones.d/master/cluster-health.conf
new file mode 100644
index 0000000..c1d8956
--- /dev/null
+++ b/zones.d/master/cluster-health.conf
@@ -0,0 +1,20 @@
+apply Service "agent-health" {
+	import "generic-service"
+	zone = "master"
+
+	check_command = "cluster-zone"
+
+	assign where host.vars.os
+	ignore where host.vars.cluster_zone == "master"
+}
+
+apply Dependency "agent-health-check" to Service {
+	parent_service_name = "agent-health"
+
+	states = [ OK ]
+	disable_notifications = true
+
+	assign where host.vars.os
+	ignore where host.vars.cluster_zone == "master"
+	ignore where service.name == "agent-health" // Avoid self dependency
+}