Describe the problem/error/question
I am setting up n8n on Azure Kubernetes Service. Instead of setting up PostgreSQL as container, I will be using Azure Postgresql Database.
I see the documentation and git repository to setup here - n8n-hosting/kubernetes at main · n8n-io/n8n-hosting · GitHub. However, I am unsure on following:
- In non-queue mode, for n8n pods, should I change replica to > 1 e.g. lets say 3? If I do so, will it mean all these 3 pods will be processing requests in parallel thus serving more executions per minute?
- If I want to serve 10 requests per minute, and if my workflows are complex e.g. go through 2000 rows of Excel and call LLMs for batch of rows, or processing PDF and extracting columns, then how much RAM should I setup for my n8n pods?
- For above load, how much RAM should I configure my Postgresql to be?
- If I use queue mode, it says filesystem is not supported, so binary data may crash, so how big images or binary data will I be able to handle if I am using queue mode?
- Is there a github deployment repo to setup queue mode workers/ui/redis using AKS?
Information on your n8n setup
- n8n version: Enterprise (yet to purchase the subscription, in planning phase)
- Database (default: SQLite): Azure Postgresql - External
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via (Docker, npm, n8n cloud, desktop app): AKS self-hosted
- Operating system: Linux AKS
Thanks and best regards,
Ashish