Connectivity diagnostic sensors (v0.2.0)#1
Conversation
Add per-device diagnostic entities (entity_category=diagnostic, auto-created, no YAML) to monitor whether an LD2450 stays online: - binary_sensor ble_connected (connectivity) - sensor connection_state (enum: connected/reconnecting/disconnected/stale) - sensor last_seen, last_disconnect (timestamp) - sensor disconnect_count, reconnect_count - sensor offline_duration, online_duration (seconds) Central state tracking in the coordinator: last_seen on each packet, one disconnect counted per interruption, reconnect counted on recovery, stale watchdog (STALE_TIMEOUT=30s), periodic refresh so durations tick while offline. Diagnostic entities stay available when the device is offline. Added LD2450BLE.is_connected. No changes to existing entity ids/behaviour. reliability_24h intentionally deferred (needs persistent history) - documented as TODO. Tests: 7 new diagnostic tests (initial/last_seen/disconnect-once/reconnect/ stale/states/durations). ruff + 35 pytest tests pass. Verified live on real HLK-LD2450 hardware (48 entities incl. 8 diagnostics).
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9b0c294562
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if self.device.is_connected: | ||
| if ( | ||
| self.last_seen is not None | ||
| and (now - self.last_seen).total_seconds() > STALE_TIMEOUT | ||
| ): | ||
| return STATE_STALE | ||
| return STATE_CONNECTED |
There was a problem hiding this comment.
When the BLE client connects but no valid target frame has ever been decoded (for example notifications are not delivered or parsing rejects all frames), last_seen remains None, so this branch returns connected indefinitely and ble_connected is true. The stale watchdog is only armed from _async_emit() after a decoded update, so the new diagnostics never transition to stale in this no-data scenario, which defeats the intended "fresh data" connectivity check.
Useful? React with 👍 / 👎.
2 participants
Adds per-device diagnostic entities (entity_category=diagnostic, auto-created, no YAML) to monitor whether an HLK-LD2450 stays reliably online.
New entities: ble_connected (connectivity), connection_state (enum), last_seen, last_disconnect, disconnect_count, reconnect_count, offline_duration, online_duration.
Implementation: central state tracking in the coordinator (stale watchdog 30s, disconnect counted once per interruption, reconnect on recovery, heartbeat refresh); diagnostics stay available when offline; new LD2450BLE.is_connected; no changes to existing entity IDs/behaviour.
Deferred: reliability_24h (needs persistent history) - documented as TODO.
Quality: ruff clean, 35 pytest tests pass (7 new); verified live on real HLK-LD2450 hardware (48 entities incl. 8 diagnostics).