Adding alertmanager support

dicos/70_prometheus.xml

@@ -1,9 +1,12 @@
 <?xml version="1.0" encoding="utf-8"?>
 <creole>
     <files>
-        <file filelist='prometheus' name='/etc/prometheus/prometheus.yml' source='prometheus.yml' mkdir='True' rm='True'/>
+        <file filelist='prometheus' name='/etc/prometheus/prometheus.yml'   mkdir='True' rm='True'/>
-        <file filelist='grafana' name='/etc/grafana/grafana.ini' source='grafana.ini' mkdir='True' rm='True'/>
+        <file filelist='prometheus' name='/etc/prometheus/alertmanager.yml' mkdir='True' rm='True'/>
+        <file filelist='grafana'    name='/etc/grafana/grafana.ini'         mkdir='True' rm='True'/>
+
         <service>prometheus</service>
+        <service>alertmanager</service>
         <service>grafana-server</service>
         <service_access service='prometheus'>
             <port service_accesslist="saLemon">80</port>
@@ -79,8 +82,71 @@
                 <value>false</value>
             </variable>
     </family>
+
+    <family name="alertes prometheus">
+        <variable name='alSMTPHost' type='string' description="Adresse du serveur SMTP pour l'envois des alertes"/>
+        <variable name='alSMTPPort' type='string' description="Port d'écoute du serveur SMTP pour l'envois des alertes"/>
+        <variable name='alFrom' type='string' description="Adresse d'origine des emails d'alerte"/>
+        <variable name='alSMTPAuth' type='oui/non' description="Authentification requise sur le serveur SMTP ?">
+            <value>non</value>
+        </variable>
+        <variable name='alSMTPUser' type='string' description="Utilisateur SMTP"/>
+        <variable name='alSMTPPass' type='string' description="Mot de passe"/>
+
+        <variable name='alDefaultReceiver' type='string' description='Nom du "receiver" par défaut'/>
+
+        <variable name='alReceiver' type='string' description="Nom du destinataire"/>
+        <variable name='alReceiverEmail' type='string' description="Adresse email du destinataire"/>
+
+        <variable name='alRoute' type='string' description="Nom de la rêgle de distribution des alertes" multi="true"/>
+        <variable name='alRouteMatchSource' type='string' description='Source de correspondance'/>
+        <variable name='alRouteMatchValue' type='string' description='Valeur attendue'/>
+        <variable name='alRouteMatchReceiver' type='string' description="Equipe destinataire de l'alerte"/>
+
+        <variable name='alRouteRegxp' type='string' description="Rêgle de distribution des alertes" multi="true"/>
+        <variable name='alRouteMatchRegExpSource' type='string' description='Source de correspondance'/>
+        <variable name='alRouteMatchRegExp' type='string' description='Expression régulière'/>
+        <variable name='alRouteMatchRegxpRecv' type='string' description="Equipe destinataire de l'alerte (regxp)"/>
+
+		<variable name='alSubRoute' type='string' description="Nom de la rêgle maitresse"/>
+        <variable name='alSubRouteMatchSource' type='string' description='Source de correspondance'/>
+        <variable name='alSubRouteMatchValue' type='string' description='Valeur attendue'/>
+        <variable name='alSubRouteMatchReceiver' type='string' description="Equipe destinataire de l'alerte"/>
+    </family>
+
+    <separators>
+        <separator name='alSMTPHost'>Configuration SMTP pour l'envois des alertes</separator>
+        <separator name='alDefaultReceiver'>Destinatires</separator>
+        <separator name='alRoute'>Rêgles de distribution</separator>
+        <separator name='alSubRoute'>Sous-rêgles de distribution</separator>
+    </separators>
+
     </variables>
     <constraints>
+
+        <group master='alReceiver'>
+            <slave>alReceiverEmail</slave>
+            <slave>prCliSonde</slave>
+        </group>
+
+        <group master='alRoute'>
+            <slave>alRouteMatchSource</slave>
+            <slave>alRouteMatchValue</slave>
+            <slave>alRouteMatchReceiver</slave>
+        </group>
+
+        <group master='alRouteRegxp'>
+            <slave>alRouteMatchRegExpSource</slave>
+            <slave>alRouteMatchRegExp</slave>
+            <slave>alRouteMatchRegxpRecv</slave>
+        </group>
+
+        <group master='alSubRoute'>
+            <slave>alSubRouteMatchSource</slave>
+            <slave>alSubRouteMatchValue</slave>
+            <slave>alSubRouteMatchReceiver</slave>
+        </group>
+
         <group master='prCli'>
             <slave>prCliIP</slave>
             <slave>prCliSonde</slave>

tmpl/alertmanager.yml

@@ -0,0 +1,119 @@
+global:
+  # The smarthost and SMTP sender used for mail notifications.
+  smtp_smarthost: '%%alSMTPHost:%%alSMTPPort'
+  smtp_from: '%%alFrom'
+%if %%getVar('alSMTPAuth','non') == 'oui'
+  smtp_auth_username: '%%alSMTPUser'
+  smtp_auth_password: 'alSMTPPass'
+%end if
+  # The auth token for Hipchat.
+  #hipchat_auth_token: '1234556789'
+  # Alternative host for Hipchat.
+  #hipchat_api_url: 'https://hipchat.foobar.org/'
+
+# The directory from which notification templates are read.
+templates: 
+- '/etc/alertmanager/template/*.tmpl'
+
+# The root route on which each incoming alert enters.
+route:
+  # The labels by which incoming alerts are grouped together. For example,
+  # multiple alerts coming in for cluster=A and alertname=LatencyHigh would
+  # be batched into a single group.
+  group_by: ['alertname', 'cluster', 'service']
+
+  # When a new group of alerts is created by an incoming alert, wait at
+  # least 'group_wait' to send the initial notification.
+  # This way ensures that you get multiple alerts for the same group that start
+  # firing shortly after another are batched together on the first 
+  # notification.
+  group_wait: 30s
+
+  # When the first notification was sent, wait 'group_interval' to send a batch
+  # of new alerts that started firing for that group.
+  group_interval: 5m
+
+  # If an alert has successfully been sent, wait 'repeat_interval' to
+  # resend them.
+  repeat_interval: 3h 
+
+  # A default receiver
+  receiver: %%alDefaultReceiver
+
+  # All the above attributes are inherited by all child routes and can 
+  # overwritten on each.
+
+  # The child route trees.
+  routes:
+  # This routes performs a regular expression match on alert labels to
+  # catch alerts that are related to a list of services.
+%for route in %%getVar('alRouteRegxp',[])
+  - match_re:
+      %%{route.alRouteMatchRegExpSource}: %%{route.alRouteMatchRegExp}
+      receiver: %%route.alRouteMatchRegxpRecv
+  %if not is_empty('alSubRoute')
+    routes:
+    %for sroute in %%getVar('alSubRoute',[])
+    # The service has a sub-route for critical alerts, any alerts
+    # that do not match, i.e. severity != critical, fall-back to the
+    # parent node and are sent to 'team-X-mails'
+      %if %%sroute == %%route
+    - match:
+        %%{sroute.alSubRouteMatchSource}: %%alSubRouteMatchValue
+      receiver: %%alSubRouteMatchReceiver
+      %end if
+    %end for
+  %end if
+%end for
+%for rt in %%getVar('alRoute',[])
+  - match:
+      %%{rt.alRouteMatchSource}: %%{rt.alRouteMatchValue}
+    receiver: %%rt.alRouteMatchReceiver
+
+  %if not is_empty('alSubRoute')
+    routes:
+    %for sroute in %%getVar('alSubRoute',[])
+      %if %%sroute == %%rt
+    - match:
+        %%{rt.alSubRouteMatchReceiver}: %%{rt.alSubRouteMatchReceiver}
+      receiver: %%rt.alSubRouteMatchReceiver
+      %end if
+    %end for
+  %end if
+%end for
+
+#  # This route handles all alerts coming from a database service. If there's
+#  # no team to handle it, it defaults to the DB team.
+#  - match:
+#      service: database
+#    receiver: team-DB-pager
+#    # Also group alerts by affected database.
+#    group_by: [alertname, cluster, database]
+#    routes:
+#    - match:
+#        owner: team-X
+#      receiver: team-X-pager
+#    - match:
+#        owner: team-Y
+#      receiver: team-Y-pager
+
+
+# Inhibition rules allow to mute a set of alerts given that another alert is
+# firing.
+# We use this to mute any warning-level notifications if the same alert is 
+# already critical.
+inhibit_rules:
+- source_match:
+    severity: 'critical'
+  target_match:
+    severity: 'warning'
+  # Apply inhibition if the alertname is the same.
+  equal: ['alertname', 'cluster', 'service']
+
+
+receivers:
+%for rcv in %%getVar('alReceiver',[])
+- name: '%%rcv'
+  email_configs:
+  - to: '%%rcv.alReceiverEmail'
+%end for

tmpl/prometheus.yml

@@ -40,6 +40,7 @@ scrape_configs:
     %end if
 ]
 %end if
+
 #alerting:
 #  alertmanagers:
 #  - scheme: https