The monitor-promdiscovery tool generate file-based service discovery files for Prometheus, for the
monitor-exporter and icinga2-exporter.
The tool is typically run from cron or equivalent tool to check OP5 Monitor or Icinga2 for hosts service that should be
scraped for performance metrics by the exporters.
For a host to be detected, it must be a part of hostgroup(s) or servicegroup(s) that is defined in the configuration
file. For a servicegroup(s) configuration all hosts that has at least on service in the configured servicegroup(s) will
be added.
Servicegroups configuration is not supported for icinga2
- monitor-promdiscovery is executed at regular intervals, e.g. every minute, by cron.
- monitor-promdiscovery performs a filter query against the configured Monitor instance for all hosts that are part of a specific hostgroup. All hosts in the hostgroup(s) will be scraped for all their services' performance data monitor-exporter or icinga2-exporter.
- monitor-promdiscovery checks against the configured file-based discovery if any host from the above call is new or not existing anymore.
- Only if there is not a match the file-based discovery file will be updated
- If the file is updated, Prometheus will reload the configuration file.
python -m monitor_promdiscovery -f ./config.yml
python -m monitor_promdiscovery -f ./config.yml -s <system>
Where can be either icinga2 or monitor
-
Check out / clone the git repo.
-
Install dependencies
pip install -r requirements.txt
-
Build a distribution
python setup.py sdist
-
Install locally
pip install dist/monitor-promdiscovery-X.Y.Z.tar.gz
The monitor-promdiscovery configuration file should be rather self-explanatory.
NB! Only icinga2 or op5monitor has to be configured. If both are used, use the -s switch instead of configure system in the property file. Using the below config executing without specifiying -s , the icinga2 entry will be used.
# The system can be either op5monitor or icinga2
system: icinga2
op5monitor:
url: https://monitor.local
user: monitor
passwd: monitor
# Connection timeout - default 5
# timeout: 5
# Verify ssl - default False
# verify: False
# The hostgroup(s) used to select hosts as targets
hostgroup:
- prometheus
# The servicegroup(s) used to select hosts as targets
# Will include all hosts that has at least on service in the configured servicegroup(s)
servicegroup:
- servicepoint
prometheus:
# path where to create the file-based discovery file - must be set
sd_file: <sd_directory>/monitor_sd.yml
# Additional labels to tag metrics with - optional
labels:
source: monitor
env: prod
icinga2:
url: https://127.0.0.1:5665
user: root
passwd: cf593406ffcfd2ef
# Connection timeout - default 5
# timeout: 5
# Verify ssl - default False
# verify: False
# The hostgroup used to select hosts as targets
hostgroup:
- prometheus
- linux_servers
prometheus:
# path where to create the file-based discovery file - must be set
sd_file: /home/andersh/programs/prometheus/sd/icinga2_sd.yml
# Additional labels to tag metrics with - optional
labels:
source: icinga2
env: prod
logger:
# Path and name for the log file. If not set sent to stdout
logfile: monitor-promdiscovery.log
# Format if day will have current day as post fix
# format: day
# Log level
#level: INFO
The hostgroup property can be both a single value or a list of different hostgroups as shown above.
The Prometheus configuration file would typical have the following job configuration in the scrape_configs section for the monitor-exporter:
- job_name: 'op5'
scrape_interval: 2m
metrics_path: /metrics
file_sd_configs:
- files:
- 'sd/monitor_sd.yml'
relabel_configs:
- source_labels: [__address__]
target_label: __param_target
- source_labels: [__param_target]
target_label: instance
- target_label: __address__
# The address for the monitor-exporter
replacement: localhost:5000
And the format for icinga2 is almost the same:
- job_name: 'icinga2'
scrape_interval: 2m
metrics_path: /metrics
file_sd_configs:
- files:
- 'sd/icinga2_sd.yml'
relabel_configs:
- source_labels: [__address__]
target_label: __param_target
- source_labels: [__param_target]
target_label: instance
- target_label: __address__
# The address for the monitor-exporter
replacement: localhost:5000
The
file_sd_configs
must have the same filename as created by the tool, as defined in the configuration file.
Python 3.6
For required packages, please review requirements.txt