Greetings,

I hope this message finds you well. I'm currently looking for some crucial information about how cAdvisor counts concurrent congestion.

I've been monitoring congestion using the following cAdvisor metrics:

• container_pressure_cpu_stalled_seconds_total
• container_pressure_memory_stalled_seconds_total
• container_pressure_io_stalled_seconds_total

To get a more accurate picture, I need to know the total congestion. My question is: if there are two concurrent congestion events, for example, on the CPU and memory, does cAdvisor count this as two separate congestion events (one for CPU and one for memory) or as a single congestion event?

If it's a single event, the total is simply the sum of the three metrics. However, if it counts them separately, I need to find a way to avoid double-counting. I need to know when there's concurrent congestion and count it only once for the total. This seems like a very complex PromQL query that is currently beyond my skill set.

Any help on this would be greatly appreciated.
Kind regards, Beladric.