We would like to ask for some new counter to be exposed by something like "ovs-appctl coverage/read-counter"

1. The too long dump duration:

• https://github.com/openvswitch/ovs/blob/bf843fd439b25e2048f8b2b466557bd3d682072d/ofproto/ofproto-dpif-upcall.c#L1052
• Logged at INFO level, we may consider promoting this to a WARNING and maybe add a coverage counter as well.
• This one indicates that revalidators are aggressively reducing the dynamic datapath flow limit.

2. Unreasonably long poll intervals.

• this is a message that is logged when the component is overloaded, potentially may be used to identify performance issues (e.g. when this message is logged at least N times during 10 minutes, N is still to be defined)
• would be nice to have a counter for this event too

more ideas to signal about scale issues are welcome