mgr/cephadm: Add PrometheusService

Signed-off-by: Sebastian Wagner <sebastian.wagner@suse.com>

diff --git a/src/pybind/mgr/cephadm/module.py b/src/pybind/mgr/cephadm/module.py
index 5c430476cd4c508eecd14f38bfc32cb84c996fc9..a54f3fc1e375b4badb2fda5e5aa570467b8c289f 100644 (file)
--- a/src/pybind/mgr/cephadm/module.py
+++ b/src/pybind/mgr/cephadm/module.py
@@ -40,7 +40,7 @@ from .services.cephadmservice import MonService, MgrService, MdsService, RgwServ
     RbdMirrorService
 from .services.nfs import NFSService
 from .services.osd import RemoveUtil, OSDRemoval, OSDService
-from .services.monitoring import GrafanaService, AlertmanagerService
+from .services.monitoring import GrafanaService, AlertmanagerService, PrometheusService
 from .inventory import Inventory, SpecStore, HostCache
 
 try:
@@ -430,6 +430,7 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule):
         self.rbd_mirror_service = RbdMirrorService(self)
         self.grafana_service = GrafanaService(self)
         self.alertmanager_service = AlertmanagerService(self)
+        self.prometheus_service = PrometheusService(self)
 
     def shutdown(self):
         self.log.debug('shutdown')
@@ -1829,7 +1830,7 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule):
         deps = []  # type: List[str]
         cephadm_config = {} # type: Dict[str, Any]
         if daemon_type == 'prometheus':
-            cephadm_config, deps = self._generate_prometheus_config()
+            cephadm_config, deps = self.prometheus_service.generate_config()
             extra_args.extend(['--config-json', '-'])
         elif daemon_type == 'grafana':
             cephadm_config, deps = self.grafana_service.generate_config()
@@ -1941,7 +1942,7 @@ class CephadmOrchestrator(orchestrator.Orchestrator, MgrModule):
             'nfs': self.nfs_service.create,
             'grafana': self.grafana_service.create,
             'alertmanager': self.alertmanager_service.create,
-            'prometheus': self._create_prometheus,
+            'prometheus': self.prometheus_service.create,
             'node-exporter': self._create_node_exporter,
             'crash': self._create_crash,
             'iscsi': self._create_iscsi,
@@ -2326,108 +2327,12 @@ api_secure = {api_secure}
     def apply_nfs(self, spec):
         return self._apply(spec)
 
-    def _generate_prometheus_config(self):
-        # type: () -> Tuple[Dict[str, Any], List[str]]
-        deps = []  # type: List[str]
-
-        # scrape mgrs
-        mgr_scrape_list = []
-        mgr_map = self.get('mgr_map')
-        port = None
-        t = mgr_map.get('services', {}).get('prometheus', None)
-        if t:
-            t = t.split('/')[2]
-            mgr_scrape_list.append(t)
-            port = '9283'
-            if ':' in t:
-                port = t.split(':')[1]
-        # scan all mgrs to generate deps and to get standbys too.
-        # assume that they are all on the same port as the active mgr.
-        for dd in self.cache.get_daemons_by_service('mgr'):
-            # we consider the mgr a dep even if the prometheus module is
-            # disabled in order to be consistent with _calc_daemon_deps().
-            deps.append(dd.name())
-            if not port:
-                continue
-            if dd.daemon_id == self.get_mgr_id():
-                continue
-            addr = self.inventory.get_addr(dd.hostname)
-            mgr_scrape_list.append(addr.split(':')[0] + ':' + port)
-
-        # scrape node exporters
-        node_configs = ''
-        for dd in self.cache.get_daemons_by_service('node-exporter'):
-            deps.append(dd.name())
-            addr = self.inventory.get_addr(dd.hostname)
-            if not node_configs:
-                node_configs = """
-  - job_name: 'node'
-    static_configs:
-"""
-            node_configs += """    - targets: {}
-      labels:
-        instance: '{}'
-""".format([addr.split(':')[0] + ':9100'],
-           dd.hostname)
-
-        # scrape alert managers
-        alertmgr_configs = ""
-        alertmgr_targets = []
-        for dd in self.cache.get_daemons_by_service('alertmanager'):
-            deps.append(dd.name())
-            addr = self.inventory.get_addr(dd.hostname)
-            alertmgr_targets.append("'{}:9093'".format(addr.split(':')[0]))
-        if alertmgr_targets:
-            alertmgr_configs = """alerting:
-  alertmanagers:
-    - scheme: http
-      path_prefix: /alertmanager
-      static_configs:
-        - targets: [{}]
-""".format(", ".join(alertmgr_targets))
-
-        # generate the prometheus configuration
-        r = {
-            'files': {
-                'prometheus.yml': """# generated by cephadm
-global:
-  scrape_interval: 5s
-  evaluation_interval: 10s
-rule_files:
-  - /etc/prometheus/alerting/*
-{alertmgr_configs}
-scrape_configs:
-  - job_name: 'ceph'
-    static_configs:
-    - targets: {mgr_scrape_list}
-      labels:
-        instance: 'ceph_cluster'
-{node_configs}
-""".format(
-    mgr_scrape_list=str(mgr_scrape_list),
-    node_configs=str(node_configs),
-    alertmgr_configs=str(alertmgr_configs)
-    ),
-            },
-        }
-
-        # include alerts, if present in the container
-        if os.path.exists(self.prometheus_alerts_path):
-            with open(self.prometheus_alerts_path, "r") as f:
-                alerts = f.read()
-            r['files']['/etc/prometheus/alerting/ceph_alerts.yml'] = alerts
-
-        return r, sorted(deps)
-
     def _get_dashboard_url(self):
         # type: () -> str
         return self.get('mgr_map').get('services', {}).get('dashboard', '')
 
     def add_prometheus(self, spec):
-        return self._add_daemon('prometheus', spec, self._create_prometheus)
-
-    def _create_prometheus(self, daemon_id, host):
-        return self._create_daemon('prometheus', daemon_id, host)
+        return self._add_daemon('prometheus', spec, self.prometheus_service.create)
 
     @trivial_completion
     def apply_prometheus(self, spec):

diff --git a/src/pybind/mgr/cephadm/services/monitoring.py b/src/pybind/mgr/cephadm/services/monitoring.py
index f05ff4fe005b05a22f3bb53eacabd7d6d5b0641a..f96ea5a61431dbe548dd1d8e1d6406011516fe0d 100644 (file)
--- a/src/pybind/mgr/cephadm/services/monitoring.py
+++ b/src/pybind/mgr/cephadm/services/monitoring.py
@@ -1,4 +1,5 @@
 import logging
+import os
 from typing import List, Any, Tuple, Dict
 
 from cephadm.services.cephadmservice import CephadmService
@@ -169,3 +170,101 @@ receivers:
             },
             "peers": peers
         }, sorted(deps)
+
+
+class PrometheusService(CephadmService):
+    def create(self, daemon_id, host) -> str:
+        return self.mgr._create_daemon('prometheus', daemon_id, host)
+
+    def generate_config(self):
+        # type: () -> Tuple[Dict[str, Any], List[str]]
+        deps = []  # type: List[str]
+
+        # scrape mgrs
+        mgr_scrape_list = []
+        mgr_map = self.mgr.get('mgr_map')
+        port = None
+        t = mgr_map.get('services', {}).get('prometheus', None)
+        if t:
+            t = t.split('/')[2]
+            mgr_scrape_list.append(t)
+            port = '9283'
+            if ':' in t:
+                port = t.split(':')[1]
+        # scan all mgrs to generate deps and to get standbys too.
+        # assume that they are all on the same port as the active mgr.
+        for dd in self.mgr.cache.get_daemons_by_service('mgr'):
+            # we consider the mgr a dep even if the prometheus module is
+            # disabled in order to be consistent with _calc_daemon_deps().
+            deps.append(dd.name())
+            if not port:
+                continue
+            if dd.daemon_id == self.mgr.get_mgr_id():
+                continue
+            addr = self.mgr.inventory.get_addr(dd.hostname)
+            mgr_scrape_list.append(addr.split(':')[0] + ':' + port)
+
+        # scrape node exporters
+        node_configs = ''
+        for dd in self.mgr.cache.get_daemons_by_service('node-exporter'):
+            deps.append(dd.name())
+            addr = self.mgr.inventory.get_addr(dd.hostname)
+            if not node_configs:
+                node_configs = """
+  - job_name: 'node'
+    static_configs:
+"""
+            node_configs += """    - targets: {}
+      labels:
+        instance: '{}'
+""".format([addr.split(':')[0] + ':9100'],
+           dd.hostname)
+
+        # scrape alert managers
+        alertmgr_configs = ""
+        alertmgr_targets = []
+        for dd in self.mgr.cache.get_daemons_by_service('alertmanager'):
+            deps.append(dd.name())
+            addr = self.mgr.inventory.get_addr(dd.hostname)
+            alertmgr_targets.append("'{}:9093'".format(addr.split(':')[0]))
+        if alertmgr_targets:
+            alertmgr_configs = """alerting:
+  alertmanagers:
+    - scheme: http
+      path_prefix: /alertmanager
+      static_configs:
+        - targets: [{}]
+""".format(", ".join(alertmgr_targets))
+
+        # generate the prometheus configuration
+        r = {
+            'files': {
+                'prometheus.yml': """# generated by cephadm
+global:
+  scrape_interval: 5s
+  evaluation_interval: 10s
+rule_files:
+  - /etc/prometheus/alerting/*
+{alertmgr_configs}
+scrape_configs:
+  - job_name: 'ceph'
+    static_configs:
+    - targets: {mgr_scrape_list}
+      labels:
+        instance: 'ceph_cluster'
+{node_configs}
+""".format(
+    mgr_scrape_list=str(mgr_scrape_list),
+    node_configs=str(node_configs),
+    alertmgr_configs=str(alertmgr_configs)
+    ),
+            },
+        }
+
+        # include alerts, if present in the container
+        if os.path.exists(self.mgr.prometheus_alerts_path):
+            with open(self.mgr.prometheus_alerts_path, "r") as f:
+                alerts = f.read()
+            r['files']['/etc/prometheus/alerting/ceph_alerts.yml'] = alerts
+
+        return r, sorted(deps)