Cumulative plot across deployment restarts

Hello everyone,

I have set my grafana dashboard to see some custom metrics of my deployments.

The problem is when the deployment restarts, instead of adding up, the value gets restarted.

The query:

watchdog_observer_files_processed_total{app=~"$watchdog"}

Underlying I am using the prometheus python client library and this metric is a Counter.

Do you know how I can merge this timeseries counter so I got it cumulative across deployment restarts.

Thanks!