glances
glances copied to clipboard
nicolargo
Monitoring network adapters
Describe the bug Glances doesn't monitor traffic of network usage inside docker
To Reproduce Steps to reproduce the behavior:
- Add a macvlan network to glances
- Exec into glances and ping a host inside the network -> shows usage
- Exec into another container in the same network, ping a host inside the network (not glances) -> shows no usage
Expected behavior I expect to see usage of the network adapter of other containers inside the monitored network
Desktop (please complete the following information):
- OS: Debian GNU/Linux 10 64bit / Linux 5.10.0-0.bpo.12-amd64
- Docker: 20.10.17
- Glances Version: Glances v3.2.4.2 with PsUtil v5.8.0
- Glances logs file:
2022-06-29 11:39:04,987 -- INFO -- Start Glances 3.2.4.2
2022-06-29 11:39:04,987 -- INFO -- CPython 3.9.9 (/usr/local/bin/python3) and psutil 5.8.0 detected
2022-06-29 11:39:04,994 -- INFO -- Read configuration file '/glances/conf/glances.conf'
2022-06-29 11:39:05,023 -- INFO -- Start GlancesWebServer mode
2022-06-29 11:39:05,245 -- WARNING -- Wifi lib is not compliant with Python 3, Wifi plugin is disabled
2022-06-29 11:39:06,815 -- INFO -- Glances RESTful API Server started on http://0.0.0.0:61208/api/3/
2022-06-29 11:39:06,818 -- INFO -- Glances Web User Interface started on http://0.0.0.0:61208/
2022-06-29 11:39:35,507 -- INFO -- Start Glances 3.2.4.2
2022-06-29 11:39:35,507 -- INFO -- CPython 3.9.9 (/usr/local/bin/python) and psutil 5.8.0 detected
- Glances test [output of glances --issue] (only available with Glances 3.1.7 or higher)
Additional context Since I use traefik I can't use network host
Glances uses the default Docker API to grab the network usage.
Is it working if you use the "docker stats" command line instead of Glances ?
docker stats doesn't show any eth ....
CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS bd6b628457ec nervous_jemison 0.00% 150.1MiB / 78.56GiB 0.19% 316MB / 7.41MB 14.5MB / 904MB 6
Hi, honestly I don't know how to use docker stats, if I do docker stats containername (of other container which uses the eth connection I'd like to monitor, I can't see any useful information, besides some network traffic, but not the actual bandwith usage.
@averon420 Just run 'docker stats'
The 6 column show the NET I/O
Then I see the traffic usage of all containers, which is not the actually bandwith. But yesI can see it.
The nitrate is compute by Glances based on the same metrics. So it should work... You should wait more or less 6 seconds after Glances start in order to see it.
Also try this to investigate:
glances --stdout docker