Here’s an interesting idea <https://www.theregister.com/2021/10/22/new_relic_codestream/>: if an app should crash during a production run, it can send telemetry data back to the developer providing some program state and pinpointing exact locations in the source code. This is done using eBPF (Extended Berkeley Packet Filter), which is able to collect information about running applications with minimal impact. You thought packet-filtering was only about managing network activity, but no, it is more useful than that. Yes, there is potential for this sort of thing to be misused, if it is not correctly managed. With great power comes great responsibility etc. By the way, while eBPF may have its roots in the older BPF as found in the BSDs, it is nevertheless a Linux-only technology <https://ebpf.io/>. Though Microsoft is trying to port some parts of it to Windows <https://cloudblogs.microsoft.com/opensource/2021/05/10/making-ebpf-work-on-windows/>.