possible bug: netbox sync stops after a few hundred elements

[FJuette]

Currently, we are testing Infrahub and the synchronization from NetBox to Infrahub (K8s deployment using Helm, 3 replicas, 3 worker).
We have a strange bug when running the sync process.
At some random point in time the sync process slows down (from > 1000 models/s to ~25 models/s) and stops with a 502 Bad Gateway error:

Restarting the sync adds a few hundred more elements and stops again:

And restarting the sync process:

The Infrahub-UI and the Infrahub-Graph-QL-Playground is responsive all the time.
The root error comes from the infrahub-sdk:

I am running the infrahub-sync-cli inside a docker container on my local machine (MacBook with latest Rancher Desktop version).
My Dockerfile:

FROM python:3.12-slim

RUN apt-get update && apt-get upgrade -y && apt-get install -y --no-install-recommends gcc libffi-dev libssl-dev && \
	pip install --upgrade pip && \
	pip install infrahub-sync pynetbox && \
	apt-get purge -y --auto-remove gcc libffi-dev libssl-dev && \
	rm -rf /var/lib/apt/lists/*
WORKDIR /sync
COPY config.yml ./netbox/config.yml
COPY schema.yml ./netbox/schema.yml
RUN infrahub-sync generate --name from-netbox --directory netbox

[BeArchiTek]

Hey @FJuette,

Could you confirm which version you are using for everything? (infrahub-sdk, infrahub-sync, and infrahub), If it was done before we release 1.3.0, were you able to try it with the latest version?

On infrahub-sync, the sync will be faster in a branch than in the default/main branch. Which method did you use?

Does the error always occured with the same "model" or with different model ? Did you try changing the default timeout on the SDK ?