Closed t-kap closed 2 years ago
hey @t-kap. interesting issue. Do you have debug logging enabled by default or was it just for testing here? If yes, that's not recommended since it's too noisy and impacts the performance.
Yes, debug was enabled for testing session only.
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.
Hi, @peimanja! Any chances to work this out?
I see there's a fresh release v1.9.3, we will try it.
After the start of upgraded exporter, in 15 minutes the timeout is back
$ time curl localhost:9531/metrics --connect-timeout 600
real 5m33.418s
user 0m0.002s
sys 0m0.032s
There is a hope. Looks like our Prometheus server requested the exporter each 30s
+ we had artifactory.timeout=15s
in exporter parameters.
Probably the exporter drowned in Prometheus requests: while it was handling a previous request, it got new ones.
For now we've got scrapes stable 15-30s
$ time curl localhost:9531/metrics --connect-timeout 600 > /dev/null
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 175k 0 175k 0 0 8552 0 --:--:-- 0:00:21 --:--:-- 44719
real 0m21.057s
user 0m0.008s
sys 0m0.008s
$ time curl localhost:9531/metrics --connect-timeout 600 > /dev/null
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 175k 0 175k 0 0 11760 0 --:--:-- 0:00:15 --:--:-- 41965
real 0m15.313s
user 0m0.005s
sys 0m0.009s
setting
1) artifactory.timeout=60s
on the exporter side
2) And such timing on Prometheus side
scrape_interval: 120s
scrape_timeout: 90s
Hope it helps.
Overview of the Issue
Immediately after restart
curl localhost:9531/metrics
response takes 25 sec. But with each next min it grows like in twice ending up in endless wait after 30 min or so.Operating system and Environment details
Centos 7
Logs
No errors in logs
regular "Registering metric messages" and "Converting size to bytes"
System is not busy
After several days there're too many open files in logs. At this time lsof -p shows about 1000 open files.
After restart lsof is like:
Any thoughts?