Last dev build broke Homepage API calls ? only 1 widget still working

[lokiiiii]

Hello everyone,

Since last update of the dev branch, it seems that API calls are kinda broken ?

I have 4 Homepage widgets right now : cpu, process, containers and memory.
While on Homepage, I can see the cpu one (first one declared in my config file) is refreshing correctly but the 3 other widgets are only refreshed once at opening and are then frozen.

Am I the only one here ?

Thanks

[nicolargo]

Hi @lokiiiii

Perhaps it's related to the commit 4361df8 done for #3033.

Do you know how does Homepage widgets call the API ?
Does it call the /api/4/all (https://github.com/nicolargo/glances/blob/develop/docs/api/restful.rst#get-all-stats) or plugin one by one ? /api/4/cpu, /api/4/mem... ?

[lokiiiii]

I asked to Homepage's dev but I would say that each widget is doing a seperated call, waiting for confirmation.
Thanks for you quick answer ! :)

[nicolargo]

I have a quick look in the Homepage Glances' plugin code and i think that all call are done separately: https://github.com/gethomepage/homepage/blob/2f48d21bfdc32e944c2d8988521f70bd289eaa2f/src/widgets/glances/widget.js#L6C28-L6C98

It should work because when i do some pure Glances test, there is no problem:

glances -C conf/glances.conf -w --disable-webui

In another terminal:

❯ curl http://0.0.0.0:61208/api/4/cpu
{"total": 0.0, "user": 0.0, "nice": 0.0, "system": 0.0, "idle": 0.0, "iowait": 0.0, "irq": 0.0, "steal": 0.0, "guest": 0.0, "ctx_switches": 0, "interrupts": 0, "soft_interrupts": 0, "syscalls": 0, "cpucore": 8, "time_since_update": 66.26712822914124, "ctx_switches_gauge": 66946749654, "ctx_switches_rate_per_sec": 0, "interrupts_gauge": 27914449423, "interrupts_rate_per_sec": 0, "soft_interrupts_gauge": 9548502030, "soft_interrupts_rate_per_sec": 0, "syscalls_gauge": 0, "syscalls_rate_per_sec": 0}

❯ curl http://0.0.0.0:61208/api/4/mem
{"total": 33093144576, "available": 10409639936, "percent": 68.5, "used": 22683504640, "free": 693043200, "active": 14244397056, "inactive": 13915897856, "buffers": 45355008, "cached": 10847420416, "shared": 676483072}

❯ curl http://0.0.0.0:61208/api/4/containers
[{"key": "name", "name": "ccs_deployer_legacy_nhe", "id": "629c8c5b3a6ef898d2b562cdaf588294437170edc8420d750df7c76131fb20c4", "status": "running", "created": "2024-04-04T08:50:49.628391999Z", "command": "tail -f /dev/null", "io": {"cumulative_ior": 1617920, "cumulative_iow": 0}, "cpu": {"total": 0.0}, "memory": {"usage": 2023424, "limit": 33093144576, "inactive_file": 49152}, "network": {"cumulative_rx": 1794598, "cumulative_tx": 126}, "io_rx": null, "io_wx": null, "cpu_percent": 0.0, "memory_percent": null, "network_rx": null, "network_tx": null, "uptime": "2 days", "image": ["docker-ccsl-virtual.repo.aes.alcatel.fr:8443/ccs_deployer:1.0.0"], "memory_usage": 2023424, "memory_inactive_file": 49152, "memory_limit": 33093144576, "engine": "docker"}, {"key": "name", "name": "ccs_deployer_nhe", "id": "2c59000024ae1427b218f7b3d32965fd943a97cd997b52a2ec1e04e2993c250c", "status": "running", "created": "2024-11-25T15:59:37.354130913Z", "command": "tail -f /dev/null", "io": {}, "cpu": {"total": 0.0021185372005044136}, "memory": {"usage": 372736, "limit": 33093144576, "inactive_file": 0}, "network": {"cumulative_rx": 1794566, "cumulative_tx": 126}, "io_rx": null, "io_wx": null, "cpu_percent": 0.0021185372005044136, "memory_percent": null, "network_rx": null, "network_tx": null, "uptime": "2 days", "image": ["docker-ccsl-gsivv-virtual.repo.aes.alcatel.fr:8443/ccs_deployer:2.0.8"], "memory_usage": 372736, "memory_inactive_file": 0, "memory_limit": 33093144576, "engine": "docker"}]
...

Just one question, in your first message, you said:

I have 4 Homepage widgets right now : cpu, process, containers and memory.

process is not a Glances plugin (processcount, processlist, programlist are).

[lokiiiii]

in Homepage's doc, I'm refering to the widget :

process: Top 5 processes based on CPU usage. Gives an overview of which processes are consuming the most resources.
(https://gethomepage.dev/widgets/services/glances/#metrics)

[nicolargo]

Ok reproduced on my side:

curl http://0.0.0.0:61208/api/4/cpu && curl http://0.0.0.0:61208/api/4/mem
==> Only cpu is refreshed

curl http://0.0.0.0:61208/api/4/cpu && sleep 1 && curl http://0.0.0.0:61208/api/4/mem
==> cpu and mem are refreshed

I will have a look asap.

Thanks for the head up @lokiiiii

[nicolargo]

@lokiiiii i just rebuild a dev container (Alpine only, Ubuntu is not available for the moment) with a patch. Can you test it ?

[lokiiiii]

@lokiiiii i just rebuild a dev container (Alpine only, Ubuntu is not available for the moment) with a patch. Can you test it ?

it seems to work just fine now ! also think it fixes the "bug" where next-serv process could use up to 450% of the process.

[nicolargo]

@lokiiiii Hi ! I just released Glances 4.4.0 with the patch, some user have update issue on HA (see #3333). Can you test this version on your side ? Thanks !

[Ryderjj89]

This is definitely not working fine in Homepage. The memory widget (and Glances itself) is reporting I have 2 GB of free memory out of 32. This is wildly incorrect. From the terminal on the machine this is on, I am using about 46%. According to Glances (and Homepage because of this), I'm using almost all of it.

Terminal:

Glances:

Homepage Memory widget:

[nicolargo]

Nope, Glances displays the "good" information:

• total: Total physical memory available.
• used: Memory used, calculated differently depending on the platform and designed for informational purposes only.
• free: Memory not being used at all (zeroed) that is readily available; note that this doesn't reflect the actual memory available (use 'available' instead).
• available: The actual amount of available memory that can be given instantly to processes that request more memory in bytes; this is calculated by summing different memory values depending on the platform (e.g. free + buffers + cached on Linux) and it is supposed to be used to monitor actual memory usage in a cross platform fashion.

You can switch from free to available memory using the Glances configuration file:

[mem]
...
# Display available memory instead of used memory
available=True
...

For a complete explanation: https://www.linuxatemyram.com/

[Ryderjj89]

I'm running this with Docker using the ubuntu-latest-full image and this changed as of this version. Is there an environment variable that can be set?

[nicolargo]

Sorry but not for the moment, it could only be done by setting the available=True in the Glances configuration file.

[Ryderjj89]

I've done that now and I see Avail in glances itself. The homepage widget is still going by Free. So likely this will also need to be fixed there too.