AWS EC2 Instance sizing

ChirpStack (v4) Setup and configuration
1

Hi all,

I am running and testing ChirpStack in a Docker Environment that’s hosted on a AWS EC2 Instance, unfortunately it’s crashed a couple times while testing and due to the AWS ‘Auto-Scaling group’ detecting the Instances bad health due to lack of memory (1GB) and deleted all of our containers with it.

I am running Multi-regions (4 x gateway bridges) and using the Postgres integration and Node red to get Gateway State.

When I don’t run the Node Red container everything seems to be stable enough with the t3 Micro instance that I’m using :

  • 2 x vCPU and 1GB of RAM
  • 3 x Gateways and a handful of Nodes
03:28:59 AM kbmemfree   kbavail kbmemused  %memused kbbuffers  kbcached  kbcommit   %commit  kbactive   kbinact   kbdirty
03:29:41 AM     65468    160500    481160     51.92        24    263616   5514848    595.07    152588    462288        72
03:30:59 AM     65468    160564    481296     51.93        24    263688   5514540    595.04    152748    461696        48
03:31:59 AM     65468    160608    481188     51.92        24    263724   5490496    592.45    152752    459828        56
03:32:41 AM     65468    160664    481132     51.92        24    263780   5514428    595.03    152752    461672       100
03:33:59 AM     65468    160744    481060     51.91        24    263852   5514840    595.07    152756    462300        76
  • So my question is what are other people using?
  • What size do I scale up to if I aim for 2000 Gateways and 4000 nodes? with some nodes sending data from every minute to once every 8 hours.

Thanks