Plugin system consists of
docker-spectator-daemon.pyperiodic collecting data fromdocker statsdocler-spectator.pyfor collecting data, retrieve average data and list of containers with collected datadocker_containers.pymunin plugin
- Python 2.6+
argparse(pip install argparse)
cd /path/to/project/location # eg. `/root/scripts` or some another
# clone repo
git clone https://github.com/dvapelnik/docker-spectatorcd docker-spectator
# start daemon
python docker-spectator-daemon.py --start
# wait some time for data collectingcd /etc/munin/plugin-conf.d
cp /path/to/project/location/docker-spectator/templates/plugin-conf.d/docker_containers_ .
cd ../plugins
# for monitoring single container
# replace [container_name] with container's name for monitoring
# replace [field_name] with field name for monitoring one of `cpu`, `mem` or `net`
ls -s /path/to/project/location/docker-spectator/docker_containers.py docker_containers_[container_name]_[field_name]
# for batch monitoring all containers
ls -s /path/to/project/location/docker-spectator/docker_containers.py docker_containers_[field_name]Warning! Batch graph monitoring available only for
cpuandmemfields not fornetwork
service munin-node restartcd /path/to/project/location/docker-spectator
cp templates/init.d/docker-spectatord /etc/init.d/
chmod +x /etc/init.d/docker-spectatordAttention! You should change daemon directory (
DAEMON_PATH) in/etc/init.d/docker-spectatordto/path/to/project/location/docker-spectator
After this you can use init.d script for start/stop daemon which collects data from docker containers like this:
service docker-spectatord restartor
/etc/init.d/docker-spectatord restartdocker-spectator-daemon.py is the daemon which exec docker-spectator.py every 10 seconds
This value is changeable. Look at config/config.py
Look at --help firstly:
python docker-spectator.py --helpusage: docker-spectator.py [-h] [-l {DEBUG,INFO,WARNING,ERROR,CRITICAL}] [-cl]
[-a {5,10,15}] [-f {cpu,mem,net}]
[-cid CONTAINER_ID] [-cn CONTAINER_NAME] [-cids]
optional arguments:
-h, --help show this help message and exit
Common:
-l {DEBUG,INFO,WARNING,ERROR,CRITICAL}, --log-level {DEBUG,INFO,WARNING,ERROR,CRITICAL}
Define log level
Collecting data:
-cl, --collect Should collect current data?
Retrieving average data:
-a {5,10,15}, --average {5,10,15}
Specify time period in minutes for retrieve average
data
-f {cpu,mem,net}, --field {cpu,mem,net}
Specify data field
-cid CONTAINER_ID, --container-id CONTAINER_ID
Specify docker container id. Container id is preferred
if specified both
-cn CONTAINER_NAME, --container-name CONTAINER_NAME
Specify docker container name. Container id is
preferred if specified both
Container:
-cids, --container-ids
Retrieve container idsYou can use docker-spectator.py for retrieve list of containers which data already collected
python docker-spectator.py --container-idsfa0a12244b9f fancy_bebe
6e1a5f381010 little_penguin
You can get collect data from active containers
python docker-spectator.py --collectYou can retrieve average collected data for container by 5, 10, 15 minutes and specify container by name or id Retrieve cpu average value by 5 minutes container specified by id:
python docker-spectator.py --average 5 --field=cpu --container-id=fa0a12244b9f0.17
Retrieve cpu average value by 10 minutes container specified by name:
python docker-spectator.py --average 10 --field=mem --container-name=little_penguin55344679
Or network. First part of result is downloaded traffic and second is uploaded. Values separated with single space
python docker-spectator.py --average 10 --field=net --container-name=little_penguin738 55644
- In this moment network graph is not stable