Merge ~jacekn/charm-prometheus-alertmanager:master into ~prometheus-charmers/charm-prometheus-alertmanager:master

[Jacek Nykis]

Add support for clustering. Switch to long config option format now that snap supports it.

[🤖 Canonical IS Merge Bot (canonical-is-mergebot)]

This merge proposal is being monitored by mergebot. Change the status to Approved to merge.

[🤖 Canonical IS Merge Bot (canonical-is-mergebot)]

Change successfully merged at revision 20fc6ba3fc267b8ce1b6dc7d8ce112ccd6058c94

diff --git a/hooks/alertmanager-cluster-relation-broken b/hooks/alertmanager-cluster-relation-broken
new file mode 120000
index 0000000..560258f
--- /dev/null
+++ b/hooks/alertmanager-cluster-relation-broken
@@ -0,0 +1 @@
+defaulthook.py
\ No newline at end of file
diff --git a/hooks/alertmanager-cluster-relation-changed b/hooks/alertmanager-cluster-relation-changed
new file mode 120000
index 0000000..560258f
--- /dev/null
+++ b/hooks/alertmanager-cluster-relation-changed
@@ -0,0 +1 @@
+defaulthook.py
\ No newline at end of file
diff --git a/hooks/alertmanager-cluster-relation-departed b/hooks/alertmanager-cluster-relation-departed
new file mode 120000
index 0000000..560258f
--- /dev/null
+++ b/hooks/alertmanager-cluster-relation-departed
@@ -0,0 +1 @@
+defaulthook.py
\ No newline at end of file
diff --git a/hooks/alertmanager-cluster-relation-joined b/hooks/alertmanager-cluster-relation-joined
new file mode 120000
index 0000000..560258f
--- /dev/null
+++ b/hooks/alertmanager-cluster-relation-joined
@@ -0,0 +1 @@
+defaulthook.py
\ No newline at end of file
diff --git a/hooks/defaulthook.py b/hooks/defaulthook.py
new file mode 100755
index 0000000..64c8489
--- /dev/null
+++ b/hooks/defaulthook.py
@@ -0,0 +1,19 @@
+#!/usr/bin/env python3
+
+# Load modules from $JUJU_CHARM_DIR/lib
+import sys
+sys.path.append('lib')
+
+from charms.layer import basic
+basic.bootstrap_charm_deps()
+basic.init_config_states()
+
+
+# This will load and run the appropriate @hook and other decorated
+# handlers from $JUJU_CHARM_DIR/reactive, $JUJU_CHARM_DIR/hooks/reactive,
+# and $JUJU_CHARM_DIR/hooks/relations.
+#
+# See https://jujucharms.com/docs/stable/authors-charm-building
+# for more information on this pattern.
+from charms.reactive import main
+main()
diff --git a/metadata.yaml b/metadata.yaml
index 6223e99..384d6ab 100644
--- a/metadata.yaml
+++ b/metadata.yaml
@@ -14,3 +14,6 @@ subordinate: false
 provides:
   alertmanager-service:
     interface: http
+peers:
+  alertmanager-cluster:
+    interface: alertmanger-cluster
diff --git a/reactive/alertmanager.py b/reactive/alertmanager.py
index 9ae82f5..6cc06c1 100644
--- a/reactive/alertmanager.py
+++ b/reactive/alertmanager.py
@@ -2,6 +2,7 @@ import yaml
 
 from charmhelpers.core import host, hookenv, unitdata
 from charmhelpers.core.templating import render
+from charmhelpers import context
 from charms.reactive import (
     when, when_not, when_any, set_state, remove_state
 )
@@ -21,6 +22,15 @@ def templates_changed(tmpl_list):
     return any_file_changed(['templates/{}'.format(x) for x in tmpl_list])
 
 
+def get_peers():
+    peers = []
+    if not context.Relations().peer:
+        return False
+    for _, data in context.Relations().peer.items():
+        peers.append('{}:9094'.format(data.get('private-address')))
+    return sorted(peers)
+
+
 @when_not('alertmanager.installed')
 def install_packages():
     hookenv.status_set('maintenance', 'Installing software')
@@ -31,13 +41,32 @@ def install_packages():
     set_state('alertmanager.do-check-reconfig')
 
 
-def runtime_args(key=None, value=None):
+def runtime_args(key=None, value=None, update=True):
     kv = unitdata.kv()
     args = kv.get('runtime_args', {})
     if key:
-        args.update({key: value})
-        kv.set('runtime_args', args)
-    args_list = ['{} {}'.format(k, v) for k, v in args.items() if v]
+        if update:
+            args.update({key: value})
+            kv.set('runtime_args', args)
+        else:
+            if not args.get(key, []):
+                cur = set([])
+            else:
+                cur = set(args.get(key))
+            new = list(cur.union((value,)))
+            args.update({key: sorted(new)})
+            kv.set('runtime_args', args)
+
+    args_list = []
+    for k, v in args.items():
+        if not v:
+            continue
+        if isinstance(v, list):
+            for i in v:
+                args_list.append('{} {}'.format(k, i))
+        else:
+            args_list.append('{} {}'.format(k, v))
+
     # sorted list is needed to avoid data_changed() false-positives
     return sorted(args_list)
 
@@ -72,7 +101,7 @@ def write_alertmanager_config_def():
     config = hookenv.config()
     port = config.get('port', '9093')
     check_ports(port)
-    runtime_args('-web.listen-address',
+    runtime_args('--web.listen-address',
                  ':{}'.format(port))
     if config.get('external_url', False):
         vars = {
@@ -81,8 +110,16 @@ def write_alertmanager_config_def():
             # alertmanager default:
             'port': port,
         }
-        runtime_args('-web.external-url',
+        runtime_args('--web.external-url',
                      config['external_url'].format(**vars))
+    if get_peers():
+        hookenv.log('Found cluster peers: {}'.format(get_peers()))
+        runtime_args('--mesh.listen-address', '0.0.0.0:9094')
+        for peer in get_peers():
+            runtime_args('--mesh.peer', peer, False)
+    else:
+        runtime_args('--mesh.listen-address', False)
+        runtime_args('--mesh.peer', False)
     args = runtime_args()
     hookenv.log('runtime_args: {}'.format(args))
     if args:
@@ -101,11 +138,6 @@ def setup_alertmanager():
     set_state('alertmanager.do-check-reconfig')
 
 
-@when('alertmanager.started')
-def check_config():
-    set_state('alertmanager.do-check-reconfig')
-
-
 @when('alertmanager.installed')
 @when_any('alertmanager.started', 'alertmanager.do-check-reconfig')
 def check_reconfig_alertmanager():
@@ -119,6 +151,9 @@ def check_reconfig_alertmanager():
         set_state('alertmanager.do-reconfig-yml')
         set_state('alertmanager.do-reconfig-def')
 
+    if data_changed('alertmanager.peers', get_peers()):
+        set_state('alertmanager.do-reconfig-def')
+
     if templates_changed([ALERTMANAGER_YML_TMPL]):
         set_state('alertmanager.do-reconfig-yml')
 
diff --git a/templates/daemon_arguments.j2 b/templates/daemon_arguments.j2
index 624290b..fcc5da1 100644
--- a/templates/daemon_arguments.j2
+++ b/templates/daemon_arguments.j2
@@ -6,18 +6,32 @@ ARGS="
 {%- for arg in args -%}
 {{ arg }} {% endfor -%}
 "
-# Alertmanager supports the following options:
-#   -config.file string
-#     	Alertmanager configuration file name. (default "/etc/prometheus/alertmanager.yml")
-#   -log.format value
-#     	If set use a syslog logger or JSON logging. Example: logger:syslog?appname=bob&local=7 or logger:stdout?json=true. Defaults to stderr.
-#   -log.level value
-#     	Only log messages with the given severity or above. Valid levels: [debug, info, warn, error, fatal]. (default info)
-#   -storage.path string
-#     	Base path for data storage. (default "/var/lib/prometheus/alertmanager/")
-#   -version
-#     	Print version information.
-#   -web.external-url string
-#     	The URL under which Alertmanager is externally reachable (for example, if Alertmanager is served via a reverse proxy). Used for generating relative and absolute links back to Alertmanager itself. If the URL has a path portion, it will be used to prefix all HTTP endpoints served by Alertmanager. If omitted, relevant URL components will be derived automatically.
-#   -web.listen-address string
-#     	Address to listen on for the web interface and API. (default ":9093")
+##############################################################################
+# WARNINIG: Due to snap confinement the following options should not be set: #
+##############################################################################
+# --config.file
+# --storage.path
+
+# Alertmanager supports the following options
+#  --data.retention=120h      How long to keep data for.
+#  --alerts.gc-interval=30m   Interval between alert GC.
+#  --log.level=info           Only log messages with the given severity or above.
+#  --web.external-url=WEB.EXTERNAL-URL
+#                             The URL under which Alertmanager is externally reachable (for example, if
+#                             Alertmanager is served via a reverse proxy). Used for generating relative and
+#                             absolute links back to Alertmanager itself. If the URL has a path portion, it will
+#                             be used to prefix all HTTP endpoints served by Alertmanager. If omitted, relevant
+#                             URL components will be derived automatically.
+#  --web.route-prefix=WEB.ROUTE-PREFIX
+#                             Prefix for the internal routes of web endpoints. Defaults to path of
+#                             --web.external-url.
+#  --web.listen-address=":9093"
+#                             Address to listen on for the web interface and API.
+#  --mesh.listen-address="0.0.0.0:6783"
+#                             mesh listen address. Pass an empty string to disable.
+#  --mesh.peer-id="52:54:00:54:c1:12"
+#                             mesh peer ID
+#  --mesh.nickname="myhost"   mesh peer nickname
+#  --mesh.password=""         password to join the peer network (empty password disables encryption)
+#  --mesh.peer=MESH.PEER ...  initial peers (may be repeated)
+#  --version                  Show application version.