Investigating root causes of device faults is difficult, particularly for deployed devices. Typically local debugging and logging subsystems are used to determine issue causes. However, these tools have limited utility in many complex scenarios. Local debugging requires physical access, lacks a similar operating environment, and is invasive by modifying operation. Logging's unstructured nature makes efficient and effect analysis difficult. Coredumps capture detailed insight into device operation at the time of the fault in a non-invasive manner. Coredumps can be used defensively to catch existing issues. This talk introduces coredumps and how to use them in practice with Zephyr. Coredump data a valueable but often underutilized debugging resource. Coredumps contain highly structured data that can be easily parsed using GDB. Coredumps are flexible regarding what data can be captured. Let your imagination run wild! Got a memory slab that needs usage monitoring? Add it to your coredump! Sensor data that causes a unique stack overflow? Add it to your coredump! Topics covered include a brief overview of Zephyr’s coredump subsystem, how to configure the subsystem for different scenarios, how to use GDB with a coredump, and some examples of coredumps in practice.