Table des matières

Installer l'agent Telegraf via Docker

Monitorer différents services spécifiques

Monitorer InfluxDB v2

Exemple de fichier telegraf.conf complet

Exemple avec : 

# Telegraf Configuration
[global_tags]
 
[agent]
  interval = "10s"
  round_interval = true
  metric_batch_size = 1000
  metric_buffer_limit = 10000
  collection_jitter = "0s"
  flush_interval = "10s"
  flush_jitter = "0s"
  precision = ""
  hostname = ""
  omit_hostname = false
  # WARNING : set to true to debug this config file !
  debug = false
 
[[outputs.influxdb_v2]]
  urls = ["http://10.0.1.10:8086"]
  ## Token for authentication.
  token = "${INFLUXDB2_TELEGRAF_TOKEN}"
  ## Organization is the name of the organization you wish to write to.
  organization = "${INFLUXDB2_INIT_ORG}"
  ## Destination bucket to write into.
  bucket = "${INFLUXDB2_INIT_BUCKET}"
  # Get all metrics except the one with "influxdb_database" tag with value equal to "oss_metrics" :
  [outputs.influxdb_v2.tagdrop]
    influxdb_database = ["*"]
 
[[outputs.influxdb_v2]]
  urls = ["http://10.0.1.30:8086"]
  # Token for authentication.
  token = "${INFLUXDB2_TELEGRAF_TOKEN}"
  # Organization is the name of the organization you wish to write to.
  organization = "${INFLUXDB2_INIT_ORG}"
  # Destination bucket to write into.
  bucket = "oss_metrics"
  # Get only metrics with tag "influxdb_database" with a value equal to "oss_metrics" :
  tagexclude = ["influxdb_database"]
  [outputs.influxdb_v2.tagpass]
    influxdb_database = ["oss_metrics"]
 
[[inputs.conntrack]]
  files = ["ip_conntrack_count", "ip_conntrack_max", "nf_conntrack_count", "nf_conntrack_max"]
  dirs = ["/host/proc/sys/net/ipv4/netfilter","/host/proc/sys/net/netfilter"]
 
[[inputs.cpu]]
  percpu = true
  totalcpu = true
  collect_cpu_time = false
  report_active = false
 
[[inputs.disk]]
  # WARNING : for root path ("/") set the directory inside Docker container (here it's "/host"). DO NOT add a trailing "/".
  mount_points = ["/host", "/host/data"]
  ignore_fs = ["tmpfs", "devtmpfs", "devfs", "iso9660", "overlay", "aufs", "squashfs"]
 
# Monitoring of instance devices
# WARNING : see distinct devices in /dev with `df -h`.
[[inputs.diskio]]
  devices = ["sda", "sdb"]
 
[[inputs.docker]]
  endpoint = "unix:///var/run/docker.sock"
  gather_services = false
  container_name_include = []
  container_name_exclude = []
  timeout = "5s"
  perdevice = false
  perdevice_include = ["cpu", "blkio", "network"]
  total_include = ["cpu", "blkio", "network"]
  docker_label_include = []
  docker_label_exclude = []
 
# Monitoring of Systemd services with help of Srvstatus scripts (inputs.systemd_status not working inside a Docker container)
[[inputs.exec]]
  commands = ["cat /opt/srvstatus/status.json"]
  timeout = "5s"
  name_override = "services_stats"
  data_format = "json"
  tag_keys = ["service"]
 
[[inputs.internal]]
 
[[inputs.interrupts]]
  cpu_as_tag = true
  [inputs.interrupts.tagdrop]
    irq = ["NET_RX", "TASKLET"]
 
[[inputs.kernel]]
 
[[inputs.linux_sysctl_fs]]
 
[[inputs.mem]]
 
[[inputs.net]]
ignore_protocol_stats = true
 
[[inputs.netstat]]
 
# Monitoring of Nginx current status
[[inputs.nginx]]
  urls = ["http://172.18.5.1:9090/nginx_status"]
  response_timeout = "5s"
 
[[inputs.nstat]]
  proc_net_netstat = "/host/proc/net/netstat"
  proc_net_snmp = "/host/proc/net/snmp"
  proc_net_snmp6 = "/host/proc/net/snmp6"
  dump_zeros = true
 
[[inputs.postgresql]]
# WARNING: ${POSTGRESQL_TELEGRAF_PWD} is not replaced here. See: https://github.com/influxdata/telegraf/issues/12458
# address = "host=172.18.5.1 user=telegraf password=${POSTGRESQL_TELEGRAF_PWD} dbname=postgres sslmode=disable"
# Need to add CONNECT right on postgres database to user telegraf.
address = "host=172.18.5.1 user=telegraf password=<password> dbname=postgres sslmode=disable"
outputaddress="postgresql-floresentinelle"
max_lifetime = "0s"
databases = ["geonature2db", "gnatlas"]
 
[[inputs.processes]]
 
# For InfluxDb metrics
[[inputs.prometheus]]
  urls = ["http://10.0.1.30:8086/metrics"]
  metric_version = 1 
  [inputs.prometheus.tags]
    influxdb_database = "oss_metrics"
 
# Monitoring of Gunicorn for GeoNature, UsersHub, TaxHub and Atlas.
[[inputs.statsd]]
  protocol = "udp"
  max_tcp_connections = 250
  tcp_keep_alive = false
  service_address = ":8125"
  delete_gauges = true
  delete_counters = true
  delete_sets = true
  delete_timings = true
  percentiles = [50.0, 90.0, 99.0, 99.9, 99.95, 100.0]
  metric_separator = "_"
  parse_data_dog_tags = false
  datadog_extensions = false
  allowed_pending_messages = 10000
  percentile_limit = 1000
 
[[inputs.swap]]
 
[[inputs.system]]
 
# WARNING: disable because we can't use systemd inside a Docker container...
# Use inputs.exec instead with Srvstatus scripts on Host.
# See: https://github.com/influxdata/telegraf/issues/7689
#[[inputs.systemd_units]]
 
# Nginx access log monitoring
# WARNING : to grant permission to read log files, add "telegraf" user to "adm" group via docker-compose.yml (group_add)
[[inputs.tail]]
  files = ["/host/var/log/nginx/access.log"]
  data_format = "grok"
  grok_timezone = "Europe/Paris"
  from_beginning = true
  name_override = "nginx_access_log"
  grok_patterns = ["%{CUSTOM_LOG_FORMAT}"]
  # Grock : https://www.elastic.co/guide/en/logstash/current/plugins-filters-grok.html
  # Grock pattern : https://github.com/elastic/logstash/blob/v1.4.2/patterns/grok-patterns
  # Ex : https://github.com/influxdata/telegraf/blob/master/plugins/parsers/grok/influx_patterns.go
  grok_custom_patterns = '''
      CUSTOM_LOG_FORMAT %{IPORHOST:client_ip} (?:%{NOTSPACE:auth}|-) (?:%{NOTSPACE:ident}|-) \[%{HTTPDATE:ts}\] "(?:%{WORD:verb} %{NOTSPACE:request}(?: HTTP/%{NUMBER:http_version:float})?|%{DATA})" %{NUMBER:resp_code} (?:%{NUMBER:resp_bytes:int}|-) %{QS:referrer} %{QS:agent} rt="?%{NUMBER:request_time:float}"? uct="(?:%{NUMBER:upstream_connect_time:float}|-)" uht="(?:%{NUMBER:upstream_header_time:float}|-)" urt="(?:%{NUMBER:upstream_response_time:float}|-)" gzr="(?:%{NUMBER:gzip_ratio:float}|-)"
    '''
 
# WARNING : only for debugging
#[[outputs.file]]
#  namepass = ["nginx_access_log"]
#  files = ["/tmp/nginx_access_log.out"]
#  influx_sort_fields = true

Tester une métrique

Problème : valeurs des métriques introuvables dans InfluxDB