Hi all,

I'm developing a prometheus exporter of Netflow data from more that 200 routers.
I use a compter flowBytes with 15 labels (application, protocol, router_name, src_network, dst_network, packets, etc.) to analyze traffic on each router.
Here flowBytes represents the total bytes of each flow, a flow is defined by all labels in the metric.

My issue is the URL with metrics is not available during the identification of application and prometheus the target as being down.

Is there a way to use multi threading with the prometheus python_client module ?

How can I purge data that are already scraped by Prometheus ?

From the exporter application can we store metrics directly from Prometheus TSDB ?

BR,