Reload loop: 'syncing entity name to panel name None' causes entity flickering

[NickBorgers]

Bug Description

The integration enters an infinite reload loop that causes entities to flicker between available and unavailable states. The logs show repeated messages:

INFO [custom_components.span_panel.coordinator] Reloading SPAN Panel integration
INFO [custom_components.span_panel.select] First update: syncing entity name to panel name 'None' for select, requesting reload
INFO [custom_components.span_panel.sensors.base] First update: syncing sensor name to panel name 'None', requesting reload
INFO [custom_components.span_panel.sensors.base] First update: syncing energy sensor name to panel name 'None', requesting reload
INFO [custom_components.span_panel.switch] First update: syncing entity name to panel name 'None' for switch, requesting reload
INFO [custom_components.span_panel.coordinator] SPAN Panel integration reload completed successfully

This cycle repeats indefinitely every ~15 seconds (scan interval).

Impact

• Entities continuously flicker to "unavailable" and back
• Dashboard displays become unusable
• Automations depending on these entities may fail

Investigation Notes

1. Network connectivity is fine - API calls complete successfully in ~0.3-2s
2. migration_mode: false did not fix the issue - the name syncing logic appears to be separate from migration
3. Panel name is None - The integration is trying to sync entity names with a panel name that is None, which triggers the reload request
4. Debug logs show successful updates followed immediately by reload requests:

   DEBUG [custom_components.span_panel.span_panel] Panel update completed successfully
   INFO [custom_components.span_panel.coordinator] SPAN Panel update cycle completed - Total: 0.959s
   INFO [custom_components.span_panel.coordinator] Reloading SPAN Panel integration
   

Environment

• Home Assistant 2026.1.2
• span_panel integration (HACS)
• span-panel-api 1.1.14 (with PR All entities offline [firmware: spanos2/r202525/05] #93 applied for network error handling)
• Two Span Panels configured via zeroconf

Configuration

Both panels have:

• migration_mode: false (changed from true, no effect)
• use_device_prefix: true
• scan_interval: 15

Expected Behavior

The integration should not enter a reload loop when the panel name is None or unset. The name syncing logic should either:

1. Skip syncing if the panel name is None
2. Use a sensible default name
3. Not trigger a reload for this condition

Related

• Testing with span-panel-api PR: Handle NetworkError exceptions in retry logic for parallel requests span-panel-api#93 (network error handling - separate issue)

🤖 Generated with Claude Code

[NickBorgers]

Additional Finding

The issue is specifically that the panel name (from the Span device's API) is None, while the device name in Home Assistant is correctly set.

Debug logs show:

DEVICE_NAME_DEBUG: Base name selected: SPAN Panel (is_simulator: False)
DEVICE_NAME_DEBUG: Received device_name parameter: SPAN Panel 2
DEVICE_NAME_DEBUG: Found existing device: SPAN Panel 2
DEVICE_NAME_DEBUG: Device already has correct name: SPAN Panel 2

But then immediately:

First update: syncing entity name to panel name 'None' for select, requesting reload

The "panel name" appears to come from the Span device's own configuration (set via the Span mobile app), not from the Home Assistant device registry. If the panel hasn't been given a custom name in the Span app, the API returns None for this field.

Questions:

1. Is this expected when the panel name isn't set in the Span app?
2. Should the integration fall back to the device name or serial number when the panel name is None?
3. Is there a configuration option to disable this name-sync behavior?

🤖 Generated with Claude Code

[NickBorgers]

Update: Panel names ARE set

The user confirmed the panels have names configured in the Span app:

• "Right - HVAC and Well"
• "Left - most of House"

So the bug is that the integration is not retrieving the panel names correctly from the API - they exist but are being returned/parsed as None.

This rules out a configuration issue on the user's side.

🤖 Generated with Claude Code

[NickBorgers]

Root Cause Identified

The reload loop is caused by sensors that don't have a circuit_id attribute (panel-level sensors like main meter, feedthrough, battery, etc.).

In sensors/base.py lines 127-133:

if hasattr(self, "circuit_id"):
    circuit = span_panel.circuits.get(getattr(self, "circuit_id", ""))
    self._previous_circuit_name = circuit.name if circuit else None
else:
    self._previous_circuit_name = None  # <-- Always None!

For sensors without a circuit_id:

1. _previous_circuit_name is always None
2. On every coordinator update, the condition if self._previous_circuit_name is None: is True
3. This triggers self.coordinator.request_reload()
4. After reload, new entity instances are created with _previous_circuit_name = None again
5. Infinite loop!

The fix should:

• Skip the name-sync logic entirely for non-circuit sensors
• Or set a sentinel value (e.g., "" or "__PANEL__") instead of None for non-circuit sensors

This is not a configuration issue - it's a logic bug in the auto-sync feature.

🤖 Generated with Claude Code

[NickBorgers]

Workaround Applied - Confirmed Working

Disabled all self.coordinator.request_reload() calls in the name-sync code. This stops the reload loop and entities are now stable.

Files patched:

• sensors/base.py (4 occurrences)
• select.py (2 occurrences)
• switch.py (2 occurrences)

Patch command:

sed -i 's/self.coordinator.request_reload()/#self.coordinator.request_reload()  # DISABLED: causes reload loop/g' \
  /config/custom_components/span_panel/sensors/base.py \
  /config/custom_components/span_panel/select.py \
  /config/custom_components/span_panel/switch.py

Why this was happening:

• SPAN Right panel had 2 circuits with no name set in the Span app (name = None)
• SPAN Left panel had all circuits named
• When circuit.name is None, the condition if self._previous_circuit_name is None: is always True
• This triggers request_reload() on every update cycle
• After reload, new entity instances are created, _previous_circuit_name is None again → infinite loop

Suggested fix for the integration:
The name-sync feature should check if circuit.name is None or falsy before attempting to sync:

if self._previous_circuit_name is None and current_circuit_name:  # Only sync if name exists
    # ... sync logic

Or better yet, use a sentinel value (not None) for "not yet synced" state.

🤖 Generated with Claude Code