Eureka J12 Ultra: Valetudo shutdown after about 4 weeks due to memory usage

[Ditti4]

Are you already using Valetudo?

Yes, just started

Your questions

I have a Eureka J12 Ultra with the newest version of Valetudo installed on it. I'm currently on vacation and wanted to check on the robot and its doings and noticed that the web interface became unreachable. Looking at the logs, this happened 2 days ago on 2026-05-07, which would be almost exactly 4 weeks since its last reboot (current uptime is 31 days).

According to the logs, Valetudo started consuming more memory than it considers to be safe:

Logs

``` [2026-05-07T00:17:02.865Z] [ERROR] Valetudo is currently taking up 172347392 bytes which is more than 1/3 of available system memory. To ensure safe robot operation, it will now shutdown. { 'process.memoryUsage()': { rss: 172638208, heapTotal: 47411200, heapUsed: 37755240, external: 25357909, arrayBuffers: 22799367 }, 'process.resourceUsage()': { userCPUTime: 2395651252000, systemCPUTime: 49213754000, maxRSS: 169152, sharedMemorySize: 0, unsharedDataSize: 0, unsharedStackSize: 0, minorPageFault: 406259795, majorPageFault: 207, swappedOut: 0, fsRead: 0, fsWrite: 0, ipcSent: 0, ipcReceived: 0, signalsCount: 0, voluntaryContextSwitches: 76512456, involuntaryContextSwitches: 1559116989 }, 'v8.getHeapStatistics()': { total_heap_size: 47411200, total_heap_size_executable: 7655424, total_physical_size: 47075328, total_available_size: 22228544, used_heap_size: 37758576, heap_size_limit: 67108864, malloced_memory: 4829164, peak_malloced_memory: 12192248, does_zap_garbage: 0, number_of_native_contexts: 1, number_of_detached_contexts: 0, total_global_handles_size: 16384, used_global_handles_size: 7488, external_memory: 25357949 }, 'v8.getHeapSpaceStatistics()': [ { space_name: 'read_only_space', space_size: 262144, space_used_size: 216848, space_available_size: 0, physical_space_size: 217088 }, { space_name: 'new_space', space_size: 524288, space_used_size: 424688, space_available_size: 82576, physical_space_size: 446464 }, { space_name: 'old_space', space_size: 33570816, space_used_size: 26507320, space_available_size: 217360, physical_space_size: 33570816 }, { space_name: 'code_space', space_size: 6815744, space_used_size: 5200000, space_available_size: 1082848, physical_space_size: 6602752 }, { space_name: 'shared_space', space_size: 0, space_used_size: 0, space_available_size: 0, physical_space_size: 0 }, { space_name: 'trusted_space', space_size: 2633728, space_used_size: 1973016, space_available_size: 566868, physical_space_size: 2633728 }, { space_name: 'new_large_object_space', space_size: 0, space_used_size: 0, space_available_size: 524288, physical_space_size: 0 }, { space_name: 'large_object_space', space_size: 2764800, space_used_size: 2676780, space_available_size: 0, physical_space_size: 2764800 }, { space_name: 'code_large_object_space', space_size: 839680, space_used_size: 771808, space_available_size: 0, physical_space_size: 839680 }, { space_name: 'shared_large_object_space', space_size: 0, space_used_size: 0, space_available_size: 0, physical_space_size: 0 }, { space_name: 'trusted_large_object_space', space_size: 0, space_used_size: 0, space_available_size: 0, physical_space_size: 0 } ], 'v8.getHeapCodeStatistics()': { code_and_metadata_size: 6675367, bytecode_and_metadata_size: 889916, external_script_source_size: 2556532, cpu_profiler_metadata_size: 0 } } [2026-05-07T00:17:02.935Z] [INFO] Valetudo shutdown in progress... [2026-05-07T00:17:03.046Z] [INFO] Disconnecting from the MQTT Broker... [2026-05-07T00:17:03.074Z] [INFO] Successfully disconnected from the MQTT Broker [2026-05-07T00:17:03.096Z] [INFO] MSmartDummycloud HTTPS server shut down [2026-05-07T00:17:03.147Z] [INFO] MSmartDummycloud MQTT client disconnected: 1775113118058926161aaaaa ```

Given the project's stance on reboots, I'm more than happy to keep the robot in this limbo state for further debugging. Any help on this is much appreciated!

[Hypfer]

Anything else interesting in the logfile? Can you attach the full thing?

I am fairly confident that whatever it is, it's not in Valetudo code, meaning that this may or may not be fixable.
It could be a nodeJS runtime thing. It could also be a in a dependency.

Since dreame and roborock reboot daily, it might be that it's an issue unrelated to midea, and it just gets masked on all the other platforms.

However, Midea also is very different, as its cloud connection is MQTT/HTTPS, so there's two new things in the stack

[Ditti4]

I've removed some what seemed like private values (MAC address, serial number, ...) - hope that's alright.

valetudo.log

In case this is relevant at all, the robot performs a daily cleanup of a single room, and it's hooked up to my Home Assistant instance using MQTT. Everything else should be pretty much unchanged from the default values.

[Hypfer]

I am planning on upgrading the valetudo runtime to Node v26, which may or many not change anything here.

If that is possible that is. That remains to be seen. So far, the binary only ever got larger

[Ditti4]

I'll keep my eyes peeled for that.

Anything else you want me to try with this robot before I reboot it? Also, do you have any ideas on how to prevent this in the future?

[Hypfer]

Nah, I don't think that there is anything that can be deduced from the system state.

As for the second question, I already answered it in detail.

[Hypfer]

I had another idea that might offer best of both worlds here. Let's see

Thanks for reporting

[Hypfer]

Starting tomorrow, the nightly version should at least keep running.

[Ditti4]

Thanks a bunch for looking into it - I'll give it a shot when I'm back home!