Conversation

Notices

1. Embed this notice
   ChatJERRYPT (jerry@infosec.exchange)'s status on Wednesday, 27-Dec-2023 07:36:50 JST ChatJERRYPT

   I should probably have checked how many grafana dashboards use influxdb before going through the pain of setting up and integrating influxdb. Back to Prometheus I go.

   • feld likes this.
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Wednesday, 27-Dec-2023 07:37:50 JST feld

     in reply to
     @jerry use VictoriaMetrics as your backend and you get both (plus an extended PromQL -- sometimes straight up fixes for stupid mistakes they made in the language)
     
     Plus its resource usage is like a tenth of Influx or Prometheus
   • Embed this notice
     ChatJERRYPT (jerry@infosec.exchange)'s status on Wednesday, 27-Dec-2023 07:39:50 JST ChatJERRYPT

     in reply to

     • feld

     @feld thanks. I’ll take a look.

   • Embed this notice
     feld (feld@bikeshed.party)'s status on Wednesday, 27-Dec-2023 07:39:50 JST feld

     in reply to
     @jerry You won't be disappointed. It's the Swiss Army knife of time series databases. Brilliant developer, very professional
   • Embed this notice
     feld (feld@bikeshed.party)'s status on Wednesday, 27-Dec-2023 07:44:26 JST feld

     in reply to
     @jerry Yes I use it for that.
     
     I haven't verified if you can also add it as a backend for Influx but I believe you can as well, and then both types of dashboards work and there's a single service you have to maintain that wears two masks 🎭
   • Embed this notice
     ChatJERRYPT (jerry@infosec.exchange)'s status on Wednesday, 27-Dec-2023 07:44:27 JST ChatJERRYPT

     in reply to

     • feld

     @feld it will work as a backend for dashboards written for Prometheus?

   • Embed this notice
     feld (feld@bikeshed.party)'s status on Wednesday, 27-Dec-2023 07:47:30 JST feld

     in reply to

     • feld
     @jerry If you're deploying Prometheus to act as a scraper on servers you can use their vmagent as a replacement with Prometheus config compatibility (minus a couple lines they tell you to remove) and you get a lower footprint service that does the same thing. Plus the binary vmagent is like 12MB vs Prometheus 100MB+ lol
   • Embed this notice
     yopp (alex@feed.yopp.me)'s status on Wednesday, 27-Dec-2023 08:30:20 JST yopp

     in reply to

     • feld

     @jerry @feld yes! It even uses pretty much the same scrape config. In most cases it just a drop-in replacement for prometheus

     feld likes this.
   • Embed this notice
     deliverator (deliverator@infosec.exchange)'s status on Wednesday, 27-Dec-2023 08:37:35 JST deliverator

     in reply to

     • feld

     @feld @jerry very happy to hear positive things about Victoria metrics! Our initial testing with it was very promising from a resource usage perspective in terms of space, CPU and RAM. Moving away from a collectd/influx stack, looking forward to being able to have more than ~3 labels in Prometheus land!

     feld likes this.