Performance issue on bulk update of devices in GUI and API

[JaysLachance]

Hello, we are experiencing a weird behavior on NetBox Community v4.5.0-Docker-3.4.2 where bulk update of 25 devices takes over 15 seconds (more devices takes even longer)

Environment:

Stack: NetBox (Docker Compose) running a custom image (netbox-custom:latest).
Hardware: VM with 62GB RAM, Xeon E5-2690 v3.
Database: Postgres 17
Troubleshooting & Fixes Applied So Far: We ruled out resource starvation and hardware emulation, but the bottleneck persists in the application write logic.
✅ Postgres Tuned: Upgraded from default config to High-Performance settings (shared_buffers=8GB, work_mem=64MB, synchronous_commit=off, random_page_cost=1.1). Memory usage confirmed active.
✅ VM CPU Latency: Switched Hypervisor CPU mode from "QEMU/Emulated" to "Host Passthrough" to fix instruction set latency (verified via lscpu).

Current Status:
Reads (GET): Instant.
Writes (POST): The 15s delay is isolated entirely to the Save operation.
Logs: Show an 11s+ gap between the POST 302 (save) and the subsequent redirect.
Has anyone ever faced that ? we have a really small database and are currently trying this out. Feels like it should be extremely fast based on the data we have so far.

PS : Updates trough API are taking roughly the same time.

[JaysLachance]

Has no one ever faced that ?!