seems to use a lot of CPU
aspiers opened this issue · 10 comments
When monitoring vaping with htop
, I see it spiking to between 60% and 90% of one CPU pretty much all the time. That sounds wrong for something which is just doing a bit of pinging. I upgraded to 0.5.0 but it didn't help. Any ideas how best to debug this?
Odd, when i run it with the example config from the documentation it's pretty much idle most of the time.
Can you paste your config?
Also how many users viewing the web ui?
It's just me using it, with only one web page viewing the UI.
probes:
- name: latency
type: std_fping
output:
- vodka
<a group>:
hosts:
- host: <a host>
name: <a host's name>
color: red
- host: <a host>
name: <a host's name>
color: blue
- host: <a host>
name: <a host's name>
color: orange
<a group>:
hosts:
- host: <a host>
name: <a host's name>
color: yellow
<a group>:
hosts:
- host: <a host>
name: <a host's name>
color: red
- host: <a host>
name: <a host's name>
color: orange
<a group>:
hosts:
- host: <a host>
name: <a host's name>
color: red
# - host: <a host>
# name: <a host's name>
# color: lime
# - host: <a host>
# name: <a host's name>
# color: blue
- host: <a host>
name: <a host's name>
color: green
- host: <a host>
name: <a host's name>
color: orange
<a group>:
hosts:
- host: <a host>
name: <a host's name>
color: coral
- host: <a host>
name: <a host's name>
color: brown
plugins:
- name: std_fping
type: fping
count: 10
interval: 3s
output:
- vodka
- name: vodka
type: vodka
logging:
version: 1
formatters:
simple:
format: '%(asctime)s - %(name)s - %(levelname)s: %(message)s'
handlers:
console:
class: logging.StreamHandler
level: DEBUG
formatter: simple
stream: ext://sys.stdout
loggers:
vodka:
level: DEBUG
handlers:
- console
data:
- type: fping
handlers:
- type: index
index: host
- type: store
container: list
limit: 7200
apps:
graphsrv:
tmpl_engine: jinja2
enabled: true
graphs:
multitarget:
id_field: host
type: multitarget
plot_y: avg
format_y: ms
smokestack:
id_field: host
type: smokestack
plot_y: avg
plugins:
- name: http
type: flask
host: <a host>
port: 7021
debug: true
static_url_path: /static
server: self
async: thread
routes:
/targets : graphsrv->targets
/graph_data :
methods:
- POST
- GET
target: graphsrv->graph_data
/graph : graphsrv->graph_view
/overview_read_file : graphsrv->overview_read_file
/: graphsrv->overview_view
Only thing i see different from your config and mine is the number of hosts so went ahead and upped those to 12 on my end and retested, and it still behaves fine.
Kind of curious what other users are seeing when they run vaping.
What distro / python version are you running ?
Could perhaps do some python profiling to see where it's spending all the time.
Yeah, it's odd. I'm on vacation right now but happy to try any profiling / debugging suggestions you have when I get back. Thanks for the reply!
@aspiers any chance to check things out again? We've been testing scaling and can't reproduce any issues.
Sorry, haven't been using this for a while, but I just fired it up again and will keep an eye out for issues. I'm running it with Python 2.7.14 on openSUSE Leap 15.0.
same here, 2 out of 3 seconds it's at > 90% (on modest dual core celeron).
checking fping running by itself produces less than 1% of CPU usage as one would expect...
[root@nexus ~]# uname -a
Linux nexus 3.10.0-862.14.4.el7.x86_64 #1 SMP Wed Sep 26 15:12:11 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
[root@nexus ~]# python --version
Python 2.7.5
noticed it using 99% of cpu when running without --debug
, with --debug
it's idle most of the time
are you guys running with or without --debug
@vegu I moved on and just created my own: https://github.com/nefilim/pinger