Standalone class for generating Nagios checkresult file, New config format, Skip checkresults for the hosts and services not known to Nagios

conf_parser.py

@@ -0,0 +1,50 @@
+#! /usr/bin/python
+
+from configobj import ConfigObj,ConfigObjError
+
+
+class ConfigParser:
+    def __init__(self):
+        self.clusters =[]
+
+    def parse (self,config_file):
+        try:
+            config = ConfigObj(config_file)
+
+            #get gmetad host information and nagios checkresult directory
+            self.gmetad_host = config.pop('gmetad_host')
+            self.gmetad_port = config.pop('gmetad_port')
+            self.force_dmax = config.pop('force_dmax')
+            self.tmax_grace = config.pop('tmax_grace')
+            self.strip_domains = config.pop('strip_domains')
+            self.nagios_result_dir = config.pop('nagios_result_dir')
+
+
+            for cluster_name in config.keys():
+                cluster_hosts = {}
+                #get hosts in the cluster
+                for host_name in config[cluster_name].keys():
+                    metrics = []
+                    #collect metric for each host in the cluster
+                    for metric_name in config[cluster_name][host_name].keys():
+                        metric_def = {}
+                        metric_def['service_name'] = config[cluster_name][host_name][metric_name]['service_name']
+                        if 'crit_above' in config[cluster_name][host_name][metric_name].keys():
+                            metric_def['crit_above'] = config[cluster_name][host_name][metric_name]['crit_above']
+                            metric_def['crit_below'] = None
+                        if 'crit_below' in config[cluster_name][host_name][metric_name].keys():
+                            metric_def['crit_below'] = config[cluster_name][host_name][metric_name]['crit_below']
+                            metric_def['crit_above'] = None
+                        if 'warn_above' in config[cluster_name][host_name][metric_name].keys():
+                            metric_def['warn_above'] = config[cluster_name][host_name][metric_name]['warn_above']
+                            metric_def['warn_below'] = None
+                        if 'warn_below' in config[cluster_name][host_name][metric_name].keys():
+                            metric_def['warn_below'] = config[cluster_name][host_name][metric_name]['warn_below']
+                            metric_def['warn_above'] = None
+                        metrics.append((metric_name,metric_def))
+                    for host in host_name.split(','):
+                        cluster_hosts.setdefault(host.lstrip(), []).append(metrics)
+                self.clusters.append((cluster_name,cluster_hosts))
+
+        except (ConfigObjError, IOError), e:
+            print 'Could not read  %s' % (e)

ganglia-nagios-bridge.conf

@@ -0,0 +1,87 @@
+gmetad_host = '127.0.0.1'
+gmetad_port = 8649
+# This overrides the DMAX attribute from all metrics in all hosts
+# If DMAX > 0 and TN > DMAX, then a metric state is considered
+# UNKNOWN and Nagios will potentially send an alert
+force_dmax = 0
+
+# Every collection group in gmond.conf defines a time_threshold
+# This value appears as TMAX in the XML.
+# The gmond process should normally send every metric again before
+# the value timer TN > TMAX.
+# If ganglia-nagios-bridge is polling a gmond collector
+# then a very small tmax_grace period (perhaps 5 seconds) is used.
+# If ganglia-nagios-bridge is polling a gmetad server then
+# tmax_grace should be set higher than the polling interval configured
+# in gmetad.
+tmax_grace = 30
+
+# Ganglia XML typically contains FQDNs for all hosts, as it obtains
+# the hostnames using reverse DNS lookups.  Nagios, on the other hand,
+# is often configured with just the hostname and no domain.  Setting
+# strip_domains = True will ensure that the domain part is stripped from
+# the hostname before passing it to Nagios.
+strip_domains = True
+
+# This is the directory where Nagios expects to read checkresults
+# submitted in batch
+nagios_result_dir = '/var/lib/nagios3/spool/checkresults'
+
+# This is where we select the metrics that we want to map from
+# Ganglia to Nagios service names
+# Any metric not matched in the configuration will be ignored and
+# not passed to Nagios.
+# Defintion for multiple clusters and their hosts to be monitored
+# along with their metrics is added in a nested format 
+#
+# Format overview :
+# cluster definiton specifying the clustername and hostnames and 
+# associated metrics to be monitored
+# 
+# can add mutliple cluster names
+# [cluster_name]
+# 	Add comma separated host name(s) and and the common metrics to be monitored
+#	[[hostname(s) separated by ,]]
+#		Metric name of the metric to be monitored for the hostnames	
+#		[[[metric name]]]
+# 		metric attributes: corresponding service name and threshold values		
+#		service_name =
+#		warn_above/below =
+#		crit_above/below =
+
+# Sample configuration
+
+[cluster_name]
+	[[host_01, host_02]] 
+		[[[proc_total]]]
+		service_name = Total Processes		
+		warn_above = 180
+		crit_above = 200
+		[[[load_one]]]
+		service_name = Current Load
+		warn_above = 0.1
+		crit_above = 0.3
+	[[host_02]]
+		[[[cpu_idle]]]
+		service_name = CPU IDLE
+		warn_above = 85
+		crit_above = 90
+		[[[disk_free]]]
+		service_name = DISK FREE
+		warn_below = 5
+		crit_below = 2
+		[[[cpu_speed]]]
+		service_name = CPU SPEED
+		warn_below = 2112
+		crit_below = 2000
+[Production1]
+	[[host3, host12]]
+		[[[disk_free]]]
+		service_name = DISK FREE
+		warn_below = 10
+		crit_below = 5
+	[[host1]]
+		[[[cpu_speed]]]
+		service_name = CPU SPEED
+		warn_below = 2000
+		crit_below = 1890