Document how to resolve `out of order sample` errors on Prometheus

[NassimBtk]

Use case

If you have many metricshub agents sending metrics to 1 Prometheus serveur and want to avoid out of order sample errors, you need to add this in your prometheus.yaml

storage:
  tsdb:
    # A 10min time window is enough because it can easily absorb retries and network delays.
    out_of_order_time_window: 10m

Documentation

We should explain how to resolve this issue in the Troubleshooting page of the MetricsHub doc.

[iguitton]

This information should be available in the quick start guides.