Gladys Assistant v4.67: New Nuki integration 🔐

News
1

Hello everyone,

I’m happy to present a new version of Gladys!

:sparkles: New features & improvements

  • :locked_with_key: New integration: Nuki You can now connect your Nuki smart locks to Gladys and include them in your scenes and dashboards. Thanks @ProtZ!
  • :house: HomeKit Updated the HAP dependency to the latest stable version for better compatibility and improved reliability with HomeKit. Thanks @bertrandda!
  • :floppy_disk: Gladys Plus DuckDB backups are now performed via a temporary connection to the database, which prevents your instances from being loaded into RAM once the backup is complete.

A big thanks to the contributors for this release :raising_hands:

Happy updating, everyone.

10 Likes
Développement intégration Nuki
Plantage container Gladys / mémoire pleine la nuit
3

I noticed that Gladys restarted every night, and apparently it was due to RAM overconsumption (4GB).

This change will probably fix my issue!

1 Like
4

Let me know if that’s not the case, we’ll take a look :slight_smile:

1 Like
5

A new release on a Friday?! :innocent:

6

FYI, I just tested the ‘update’ button and


1000011711
10000117111080×2388 270 KB

It doesn’t work


But no stress, it’s not vital :wink:

Edit : After several attempts to find the lines in the logs, it worked.

7

Congrats on this new integration @ProtZ!
I just performed the update in « forced Â» mode and my Gladys crashed.
I restarted the docker manually and I saw some messages regarding Nuki:

2026-01-23T22:22:09+0100 <info> index.js:15 (Object.start) Starting nuki service
2026-01-23T22:22:09+0100 <info> nuki.http.connect.js:10 (NukiHTTPHandler.connect) Nuki : Test http connection to Nuki Web API
  <rejected> Error: API Token required! Please check the arguments!
2026-01-23T22:22:09+0100 <error> index.js:15 (process.<anonymous>) unhandledRejection catched: Promise {
      at new Nuki (/src/server/services/nuki/node_modules/nuki-web-api/lib/Nuki.js:30:11)
      at NukiHTTPHandler.connect (/src/server/services/nuki/lib/http/nuki.http.connect.js:12:18)
}
2026-01-23T22:22:09+0100 <error> index.js:16 (process.<anonymous>) Error: API Token required! Please check the arguments!
    at new Nuki (/src/server/services/nuki/node_modules/nuki-web-api/lib/Nuki.js:30:11)
    at NukiHTTPHandler.connect (/src/server/services/nuki/lib/http/nuki.http.connect.js:12:18)

I’m wondering about these messages because I hadn’t enabled anything since I hadn’t received the update yet.
Now that Gladys is operational, I can see in the integration that Nuki is not enabled:

image
image2384×748 83.8 KB

Personally I don’t have a Nuki lock so I don’t need the integration (for now).
Is it normal for the Nuki service to start and request a token on Gladys startup?

EDIT: and I have a bunch of strange things with MQTT:

image
image1948×1024 217 KB

image
image1822×988 219 KB

The first lines with the numbers are my Z-Wave devices, then virtual MQTT devices, then z2m 

However I don’t see anything appearing with MQTT Explorer.

Another point: even if the service is not enabled, you can still start a search, same for the web:

image
image2390×790 57.3 KB

image
image2376×538 55.1 KB

8

I only do it if I’m available over the weekend afterwards to handle support and possible fixes :wink:

That’s the case this weekend!

On the Gladys side, we launch Watchtower with --run-once. Normally it’s supposed to work directly, but maybe something didn’t work on your side. If we observe this phenomenon from time to time we can report it to Watchtower!

Really crashed? Did Gladys restart?

Normally a failed start of an integration can’t crash Gladys, unless the crash is in a native C++ package, which I don’t think is the case for Nuki.

That surprises me :slight_smile:

That’s normal, however I agree that the service could display a clearer message in the case the user doesn’t intend to use the integration.

@ProtZ for info, normally services should throw new ServiceNotConfiguredError('') if they are not configured, which displays a clear message in the logs. It’s my fault, I hadn’t noticed that it was missing in your PR :smiley:

Exemple :

2026-01-23T16:02:51+0100 \u003cinfo\u003e service.start.js:40 (Service.start) Service nextcloud-talk is not configured, so it was not started.

Could you add it to the Nuki integration?

9

Gladys restarted after the update but no web access and the backend didn’t start correctly because usually I still have the scenes running, and here nothing.
Of course I didn’t analyze the logs at that moment :disappointed_face: but only after my manual restart.

10

Normally, you still have the logs available because a docker restart doesn’t remove the previous logs, unless you removed the container and recreated a new one :slight_smile:

11

Hello, sorry, I’ll do it as soon as possible (probably Sunday evening)

1 Like
12

:wink: big cleanup!

1 Like
13

@ProtZ Thank you very much for your corrective PR, I left you a review :slight_smile:

There is just one small thing:

14

Thanks for the fixes @ProtZ, I just approved the PR and merged it, it will be included in the next Gladys release :slight_smile:

1 Like