I love smokeping for finding jitter and latency but always wanted something that went deeper and grabbed all the related signals, so I built JustRebootIt. JRI runs a bunch of parallel pings & traceroutes, detects when a spike is happening, runs more detailed probes during the event, then can feed all that into an LLM for diagnosis.
It also reads from my UDM to gather CPU/memory/etc from the device in the event that correlates.
It also reads from my UDM to gather CPU/memory/etc from the device in the event that correlates.
Feedback and PRs welcome!