Librenms pushgateway very slow

hi i know this is alpha but still im trying, maybe someone came across this problem and can assist me.
the problem is that since i enabled the pushgateway integration each polling takes forever.
for example processors module:

Runtime for poller module ‘processors’: 7.7796 seconds with 93896 bytes
SNMP: [1/0.01s] MySQL: [3/0.01s] Prometheus: [6/9.47s] RRD: [12/0.00s]

Unload poller module processors

the push gateway runs on k8s.
any ideas?

I have the same issue… and really need this working as well.
As soon as I enable pushgateway my polling time goes up quite a bit, now granted I am polling every 2m maybe… but it isn’t like the host running librenms is starving for reasources.

Curious if you’ve made any progress on this, also if there is a way to filter what goes to pushgateway

You can see the difference when pushgateway is running, high polling time and when it is turned off.

My librenms box is in no way experiencing high load… I’ve tried turning off certain modules to see if that helps.

Pushgateway integration using http to send/receive the data, which is very slow and not well suited to this kind of usage.
its not hardware related(we used nvme gen10 servers) with 20G(bundled) NICs.
we are now using influxdb as DB to store these metrics and it is working pretty good. my suggestion is to leave the pushgeateway(which is unsupported as well) and move to influx.

Sorry since I’m not familiar but you’re in turn getting this metrics into prometheus and using grafana using influxDB instead of pushgateway? Is there any docs on this?

Found the docs. Thank You!

I’m having the same kind of issues. I’ve been struggling with that these last days, there’s no way to make it work better. We want to expose metrics to prometheus for around ~150 devices and it’s impossible. It’s even slow for a short list of devices. No problem of resources either.