Entity unavailable in Home Assistant – how to find the cause

The state unavailable does not identify the fault. It says that an existing entity cannot currently supply data. The cause may be the device, network, manufacturer’s service, MQTT broker, integration or a source entity.

Effective troubleshooting begins by establishing the scope and checking one layer at a time. This guide follows that process without deleting configuration or restarting everything blindly.

unavailable means the entity exists in Home Assistant, but the system cannot currently reach it or obtain its data. The device may be offline, its integration may have failed to load, or communication may have been interrupted.

unknown is different: the entity exists, but Home Assistant does not know its value at that moment. It may, for example, be waiting for its first reading. A missing entity, an unknown value and a communication failure are not the same problem.

Entity, device and integration

An entity is one data point or control, such as a temperature or switch. A device groups entities belonging to one piece of hardware. An integration is the communication layer between Home Assistant and a device or service. A device can therefore work partly while one of its entities is unavailable.

• One entity: check whether it was disabled, removed by the device or depends on a separate source.
• Every entity of one device: suspect power, device connectivity or its integration entry.
• Every device in one integration: check configuration, authentication, broker, coordinator or the vendor service.
• Several unrelated devices: check the network, router, DNS, VLAN, infrastructure power and Home Assistant itself.

Record when the fault began: after a restart, update, network change, password change or power cut. That detail often narrows the cause faster than another restart.

1. Open the entity details and follow the link to its device. Note the integration name and entity ID.
2. Go to Settings → Devices & services. Open the relevant integration and look for configuration or connection errors.
3. Check Settings → System → Repairs for issues Home Assistant has detected and asks you to resolve.
4. In Settings → System → Logs, search for the integration, device or entity ID. Concentrate on entries from when the problem began.
5. If Home Assistant requests reauthentication, complete that flow without removing the integration.
6. If the integration offers reload and the fault is confined to it, reload only that integration after saving useful log information.
7. A Home Assistant restart is reasonable after an update, a configuration change requiring it, or a stuck integration. Repeating restarts without reading the logs can hide the evidence.

For a wider explanation of communication methods, see Home Assistant integrations.

Confirm that the device has power and remains connected in the router. If the integration uses an IP address, compare it with the device’s current address. A DHCP reservation in the router reduces address changes after restarts.

Test the device from the same network and check its app or local interface when available. If it works there but Home Assistant cannot reach it, examine VLAN rules, firewall policy, DNS, Wi-Fi client isolation and blocked local traffic. This separates a device failure from a broken communication path.

A ping response alone does not prove that the integration works; the required port or protocol may still be blocked.

Check Home Assistant’s internet access and whether the manufacturer reports an outage. If the official app cannot retrieve data either, the fault is likely outside Home Assistant.

Common causes include an expired session or token, changed password, new login requirement, API limit or server error. Use Home Assistant’s reauthentication prompt when it appears. Do not remove the entire integration merely to sign in again, as that may affect entity IDs and dependencies.

Check Home Assistant’s connection to the broker separately from the device’s connection to the broker. A working MQTT integration does not prove that a particular device is publishing.

• Listen to the correct state topic and confirm that the device sends current data.
• If the entity uses an availability_topic, inspect its online/offline payloads and the device’s Last Will.
• After a restart, an entity may wait for a new message. Retained messages can deliver the last state to a new subscriber, but decide deliberately which data should be retained.
• No current value does not always mean the device is unavailable. Entity configuration, availability messages and received state data determine the result.

This is focused troubleshooting, not a full MQTT setup guide. The integrations guide explains the wider context.

For one device, check its battery or power, last communication and route to nearby Zigbee routers. Battery devices may sleep and report less often, so a missing map link alone does not prove a failure.

If many devices disappear together, inspect the coordinator, its Home Assistant connection and powered mesh routers. After a restart or coordinator change, allow communication to recover instead of immediately pairing everything again.

A derived entity can become unavailable even though its own configuration did not change. One unavailable source is enough. Check every entity used by the template and its current state.

Handle unknown and unavailable before calculations; the official has_value() function checks whether an entity has a usable state. Do not automatically turn missing data into zero when zero could trigger heating, charging, an alarm or another incorrect action.

Define safe behaviour when data disappears. See Home Assistant automations for the wider automation principles.

1. Open the relevant integration under Settings → Devices & services.
2. From its three-dot menu, enable debug logging for that integration only.
3. Reproduce the fault once and note the time.
4. Disable debug logging; Home Assistant prepares the log for download.
5. If offered, download diagnostics from the integration or device menu.
6. Before sharing, open the file and check for names, addresses, identifiers or other sensitive details. Integrations should redact secrets, but you should still verify the file.

Debug logging produces many entries, so enable it only for a short, controlled test and then turn it off.

• Do not remove an integration without recording its configuration and dependencies.
• Do not factory-reset a device before locating where communication fails.
• Do not restore a full backup because of one entity.
• Do not change several things at once; check the result after each test.
• Do not restart repeatedly without collecting logs, because you may hide the original error.

If it is unclear whether the device, network, integration, MQTT or Home Assistant is causing the problem, describe the symptoms. I will first check whether I can genuinely help and what level of troubleshooting is needed.

This guide was checked on 11 August 2026 against the official Home Assistant documentation:

• entity states and templates,
• troubleshooting, debug logging and logs,
• integration diagnostics,
• Repairs,
• MQTT and ZHA.