Messaging recovery through to device control
Check whether UMS and App clients resubscribe and complete control/state readback after broker changes, foreground reconnection and offline recovery. Server traffic, handshakes and tests do not by themselves prove that loop.
Timeline
UMS and App foreground reconnect/offline recovery adopted a common v2 connection path. Twelve tests and repository pushes were reported. Actual control still failed after release. Removing an old IPv6 DNS record restored the public WebSocket handshake, while local DNS caching, browser subscription and device command/readback remained unconfirmed.
Daily entry · Recovering services while checking the control path ↗Mini-program logs briefly showed a successful connection/subscription and some device/rule pages opened. A later test repeatedly received 503 from the connection API. This did not establish stable reconnect or complete device execution/state readback.
Daily entry · Mini-program checks and iOS release preparation ↗Rebuilding and clearing caches enabled mini-program login, device/temperature views and continuous live data, although developer warnings/errors remained. Local fixes to privacy-page messaging recovery and overlapping threshold labels were reportedly committed but unpushed. Receiving data did not settle command execution, state readback or full offline recovery.
Daily entry · Live telemetry and application submission reports ↗