Glances Containers API Memory usage does not match Primary Output

[gflip220]

While working to setup integration between Glances and GetHomepage.dev, I have noticed an issue where the Glances API seems to report incorrect docker container memory usage, while the Glances primary output reports correct usage. Screenshot below showing Docker Stats matching with Glances output, while Glances API reports a massively higher number, 3GB in this example. Other containers captured in this example are also have similar dependencies.

I am running this on Ubuntu 24.04 with the ubuntu-latest-full version of Glances on Docker.

Anything I should look at to troubleshoot this? Thank you!

[nicolargo]

Hi @gflip220

What's your Glances version ?

Last stable version return the following JSON object:

{'command': 'tail -f /dev/null',
  'cpu': {'total': 0.0},
  'cpu_percent': 0.0,
  'created': '2024-04-04T08:50:49.628391999Z',
  'engine': 'docker',
  'id': '629c8c5b3a6ef898d2b562cdaf588294437170edc8420d750df7c76131fb20c4',
  'image': ['docker-ccsl-virtual.repo.aes.alcatel.fr:8443/ccs_deployer:1.0.0'],
  'io': {},
  'io_rx': None,
  'io_wx': None,
  'key': 'name',
  'memory': {'inactive_file': 0, 'limit': 33096175616, 'usage': 397312},
  'memory_percent': None,
  'memory_usage': 397312,
  'name': 'ccs_deployer_legacy_nhe',
  'network': {'cumulative_rx': 11466973,
              'cumulative_tx': 0,
              'rx': 0,
              'time_since_update': 3,
              'tx': 0},
  'network_rx': 0,
  'network_tx': 0,
  'status': 'running',
  'uptime': '2 weeks'}

memory_usage field should be align with the one displayed in UI.

[gflip220]

Hello @nicolargo

I am running the current ubuntu-latest-full docker pull. Glances 4.3.0.7 with psutil 6.1.1

I can currently reproduce the issue if there are any logs or other information you would like.

Thank you!