90k executions a month?? · Brendan's AI Community

90k executions a month??

Hi team and Brendan, I need some advise on running sms agents on scale:

I will be creating an n8n agent that sends sms (through Twillio) and is connected to Salesforce to sell products to incoming leads. It is a large company that receives 500 leads a day. So about 3000 executions a day (6 messages average) ->

90k executions a month minimum.

I got:

Hostinger VPS; KVM 2 running n8n. But will probably upgrade to VPS: "n8n (queue mode)" with a KVM 8.
I have been using Qdrant as a vector base, but as this agent will only need a small to maybe no knowledge base, what should I be using on scale? What is price efficient?
GPT-5 mini.
Simple Memory: based on phone number. The chats do not have to be stored in memory, as this agent is just used as a single sales conversation. I do store the transcripts in Airtable for feedback

I don't have experience in running n8n in Hostinger VPS, nor with a (queue mode). I am also not too familiar with vector bases. I am trying to keep the workflow and agent simple, but I need to make sure it will operate at scale.

If you have any tips/advice/experience on any of these topics, please share!

Thanks guys, Dirk

1 comment