troubleshooting: add a bpf map for unexpected errors in bpf code

[kkourt]

There are many places in our bpf code where we exit without generating the event when unexpected error happens. For troubleshooting and debugging it would be good to track this information so that we know if we hit any of those unexpected error paths. This can be implemented with a BPF map that maintains a counter per code location where we return early.

[jrfastab]

Many maps exist now for individual sensors. Is this to capture cases that don't already have maps or to somehow unify the error maps?

[kkourt]

The purpose of this issue is to account for cases where we silently return. It is not about unifying the error maps.