Class: Sensu::Plugins::Prometheus::Metrics

Inherits:
Object
  • Object
show all
Includes:
Utils::Log
Defined in:
lib/sensu/plugins/prometheus/metrics.rb

Overview

Represents the queryies that will be fired against Prometheus to collect information about monitored resources. Ideally all public methods on this class will be available as a final check.

Instance Method Summary collapse

Methods included from Utils::Log

#log, log

Constructor Details

#initialize(prometheus_client) ⇒ Metrics

Returns a new instance of Metrics.



12
13
14
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 12

def initialize(prometheus_client)
  @client = prometheus_client
end

Instance Method Details

#custom(cfg) ⇒ Object

Execute query informed on check’s configuration and makes no modifications on value.



18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 18

def custom(cfg)
  metrics = []
  @client.query(cfg['query']).each do |result|
    source = if result['metric']['instance'] =~ /^\d+/
               result['metric']['app']
             else
               result['metric']['instance']
             end

    metrics << {
      'source' => source,
      'value' => result['value'][1]
    }
  end
  metrics
end

#disk(cfg) ⇒ Object

Query percentage of mountpoint total disk space size compared with avaiable.



37
38
39
40
41
42
43
44
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 37

def disk(cfg)
  mountpoint = "mountpoint=\"#{cfg['mount']}\""
  query = @client.percent_query_free(
    "node_filesystem_size{#{mountpoint}}",
    "node_filesystem_avail{#{mountpoint}}"
  )
  prepare_metrics('disk', @client.query(query))
end

#disk_all(cfg) ⇒ Object

Query percentage of free space on file-systems, ignoring by default ‘tmpfs` or the regexp configured on check.



48
49
50
51
52
53
54
55
56
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 48

def disk_all(cfg)
  ignored = cfg['ignore_fs'] || 'tmpfs'
  ignore_fs = "fstype!~\"#{ignored}\""
  query = @client.percent_query_free(
    "node_filesystem_files{#{ignore_fs}}",
    "node_filesystem_files_free{#{ignore_fs}}"
  )
  prepare_metrics('disk_all', @client.query(query))
end

#inode(cfg) ⇒ Object

Queyr percentage of free inodes on check’s configured mountpoint.



59
60
61
62
63
64
65
66
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 59

def inode(cfg)
  mountpoint = "mountpoint=\"#{cfg['mount']}\""
  query = @client.percent_query_free(
    "node_filesystem_files{#{mountpoint}}",
    "node_filesystem_files_free{#{mountpoint}}"
  )
  prepare_metrics('inode', @client.query(query))
end

#load_per_cluster(cfg) ⇒ Object

Calculates the load of an entire cluster.



122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 122

def load_per_cluster(cfg)
  cluster = cfg['cluster']
  query = format(
    'sum(node_load5{job="%s"})/count(node_cpu{mode="system",job="%s"})',
    cluster,
    cluster
  )

  result = @client.query(query).first
  source = cfg['source']
  value = result['value'][1].to_f.round(2)
  log.debug(
    "[load_per_cluster] value: '#{value}', source: '#{source}'"
  )

  [{ 'source' => source, 'value' => value }]
end

#load_per_cluster_minus_n(cfg) ⇒ Object

Returns a single metric entry, with the sum of the total load on cluster divided by the total amount of CPUs.



142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 142

def load_per_cluster_minus_n(cfg)
  cluster = cfg['cluster']
  minus_n = cfg['minus_n']
  sum_load = "sum(node_load5{job=\"#{cluster}\"})"
  total_cpus = "count(node_cpu{mode=\"system\",job=\"#{cluster}\"})"
  total_nodes = "count(node_load5{job=\"#{cluster}\"})"

  query = format(
    '%s/(%s-(%s/%s)*%d)',
    sum_load, total_cpus, total_cpus, total_nodes, minus_n
  )
  result = @client.query(query).first
  value = result['value'][1].to_f.round(2)
  source = cfg['source']
  log.debug(
    "[load_per_cluster_minus_n] value: '#{value}', source: '#{source}'"
  )

  [{ 'source' => source, 'value' => value }]
end

#load_per_cpu(_) ⇒ Object

Current load per CPU.



164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 164

def load_per_cpu(_)
  cpu_per_source = {}
  @client.query(
    '(count(node_cpu{mode="system"})by(instance))'
  ).each do |result|
    source = result['metric']['instance']
    cpu_per_source[source] = result['value'][1]
  end

  metrics = []
  @client.query('node_load5').each do |result|
    source = result['metric']['instance']
    value = result['value'][1].to_f.round(2)
    load_on_cpu = value / cpu_per_source[source].to_f
    log.debug(
      "[load_per_cpu] value: '#{load_on_cpu}', source: '#{source}'"
    )
    metrics << {
      'source' => source,
      'value' => load_on_cpu
    }
  end
  metrics
end

#memory(_) ⇒ Object

Query the percentage free memory.



95
96
97
98
99
100
101
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 95

def memory(_)
  query = @client.percent_query_free(
    'node_memory_MemTotal',
    'node_memory_MemAvailable'
  )
  prepare_metrics('memory', @client.query(query))
end

#memory_per_cluster(cfg) ⇒ Object

Percentage free memory cluster wide.



104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 104

def memory_per_cluster(cfg)
  cluster = cfg['cluster']
  query = @client.percent_query_free(
    "sum(node_memory_MemTotal{job=\"#{cluster}\"})",
    "sum(node_memory_MemAvailable{job=\"#{cluster}\"})"
  )

  metrics = []
  source = cfg['source']
  @client.query(query).each do |result|
    value = result['value'][1].to_f.round(2)
    log.debug("[memory_per_cluster] value: '#{value}', source: '#{source}'")
    metrics << { 'source' => source, 'value' => value }
  end
  metrics
end

#predict_disk_all(cfg) ⇒ Object

Compose query to predict disk usage on the last day.



69
70
71
72
73
74
75
76
77
78
79
80
81
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 69

def predict_disk_all(cfg)
  days = cfg['days'].to_i
  days_in_seconds = days.to_i * 86_400
  filter = cfg['filter'] || {}
  range_vector = cfg['sample_size'] || '24h'
  query = format(
    'predict_linear(node_filesystem_avail%s[%s], %i) < 0',
    filter,
    range_vector,
    days_in_seconds
  )
  prepare_metrics('predict_disk_all', @client.query(query))
end

#service(cfg) ⇒ Object

Service metrics will contain it’s “state” as “value”.



84
85
86
87
88
89
90
91
92
 
# File 'lib/sensu/plugins/prometheus/metrics.rb', line 84

def service(cfg)
  defaults = { 'state' => 'active' }
  cfg = defaults.merge(cfg)
  query = format(
    "node_systemd_unit_state{name='%s',state='%s'}",
    cfg['name'], cfg['state']
  )
  prepare_metrics('service', @client.query(query))
end