Help configuring Vector with HTTP sink for Azure Virtual Desktop telemetry collection

[louis030195]

Question

I'm trying to set up Vector v0.41.1 on Azure Virtual Desktop (AVD) VMs to collect Windows UI automation telemetry and forward it to Supabase via HTTP sink.

Current Setup:

• Running Vector on Windows Server in Azure VMSS (Virtual Machine Scale Set)
• Need to collect logs from MCP (Model Context Protocol) agent running UI automation with Opentelemetry integration
• Attempting to send telemetry to Supabase PostgreSQL via REST API

Issues:

1. Vector binary fails to start in Windows user session (Session 1) via scheduled task
2. Getting PowerShell syntax errors when trying to escape quotes in automated deployment scripts
3. Network connectivity issues downloading Vector from GitHub releases during VM provisioning

Has anyone successfully deployed Vector on Windows VMs with automated provisioning? Looking for examples of:

• Proper Windows scheduled task configuration for Vector
• Working HTTP sink configuration for REST APIs
• Best practices for escaping/handling Vector config in PowerShell scripts

Any help or working examples would be greatly appreciated!

Vector Config

[sources.mcp_logs]
type = "file"
include = ["C:\\MCP\\logs\\*.log"]
read_from = "beginning"

[sinks.supabase]
type = "http"
inputs = ["mcp_logs"]
uri = "https://[supabase-project].supabase.co/rest/v1/mcp_telemetry"
method = "POST"
encoding.codec = "json"
headers.apikey = "[service-role-key]"
headers.Content-Type = "application/json"

Vector Logs

No response

[pront]

Hi @louis030195, can you share how you are trying to install Vector?
You can download the Windows version from our downloads page: https://vector.dev/download/

[louis030195]

i switched to opentelemetry collector