Connectivity's back now (Slack 200, Granola 401 = reachable, just an auth header check on that bare probe — the job sends the key). The

ConnectionRefused

errors were during the offline/battery-dead window and have since recovered. Here's the actual state, Royce: Sleep cycle: Not missed — the dream cycle ran clean at 04:19 this morning. Machine was back up by then. What did get hit was the ingest during the dead window: • Granola ingest — last ran 22:16 last night, failed with ConnectionRefused. So any meeting notes from after ~10pm Fri through when the machine died aren't in yet. • Slack archive — the 01:30 run partially recovered (archived 65 messages, committed + synced) but hit a ConnectionRefused mid-run on

users.list

, so it may be incomplete. Both jobs are scheduled (Granola every 15 min, Slack nightly), so they'll self-heal on the next tick now that the network's back. Want me to force-run both now to backfill the gap immediately rather than wait for the next scheduled tick?