HENK

“Someone has to keep an eye on things.”

What he watches

The homelab: which services are up, which are not, what the sensors say. He reads from a small, closed set of tools and nothing else. Ask him “is everything up?” over Signal and he will tell you.

When something breaks

He opens one conversation, not a stream: a diagnosis with how sure he is, a suggested fix, and where to pick the work up. Storms collapse into one message. Chronic grumbles go to the record, not to you.

What he never does without asking

Change anything. Every write declares its tier in the code; anything not standing waits for an inline yes from the owner and fails closed on silence. Strangers get nothing, not even a read receipt.

Something Henk sent, (trimmed)

Recurrence — same alert as this morning. rp2's node_exporter recovered in between but dropped again with the same "connection refused" signature. Diagnosis: node_exporter process on rp2 is repeatedly crashing/stopping (host itself is fine). Root cause still unconfirmed. (confidence: moderate) Fix: SSH to rp2, restart node_exporter, but this time also check dmesg/syslog for OOM-kills or SD-card I/O errors — two occurrences in 6h means restart-and-wait isn't enough anymore. Pickup: handoff 7sKd2PqLx9Wm — a third recurrence should be treated as a standing fault.

He runs on one Raspberry Pi in a cupboard. The whole deployment is docker compose up -d.