Welcome — what evidence disappears after your device reboots?

[TheusHen]

Latch is opening its public-preview phase around one question: when your embedded device reboots after a failure, what evidence do you wish had survived?

Share the failure mode you deal with, along with the board/MCU, RTOS or bare-metal environment, compiler, reset path, and storage constraints. Brownouts, watchdog loops, corrupt stacks, field-only faults, retained-RAM surprises, and interrupted Flash writes are especially useful context.

Good places to begin:

• Try the host capture-and-decode demo in PR #10 while the public-launch baseline is under review.
• Pick a bounded good first issue.
• Browse help wanted if you can test FreeRTOS, Zephyr, or real hardware.
• Use Q&A for integration help and Ideas for protocol or architecture proposals.

Please remove device identifiers, production keys, proprietary firmware, endpoints, and captured customer memory. Security-sensitive findings belong in private vulnerability reporting, not a public discussion.

The most helpful early feedback is concrete: what failed, what disappeared at reset, what storage survives, and what a useful post-reboot envelope would need to contain.