Are my soil moisture alarm edits and / or use of the graphs hanging the server?

[jamiem...@gmail.com]

Server hung again... as of 5:55 eastern on 7/21/2025.

Ever since configuring low soil moisture on my 10 sensors several days ago, I get a lot of UI problems, that end up with The Service is unavailable. 

Unable Probably 4 times over the past 10 days I log into my account via chrome desktop browser, and either change moisture alarm settings or just look at my various soil moisture readings of my various plants, and the errors below arise, until I cannot get into the site at all! Just happened right now!

Today I got the same / similar yellow banner after using the site a few minutes.  

Unknown Error: {"readyState":4,"responseText":"The service is unavailable.","status":503,"statusText":"Service Unavailable"]

First only on one sensor (called Powder Puff) did this while a couple others still worked,  then I seem to get this repeatedly, or just see very long delays. When I refreshed  the page https://www.mytaglist.com/eth/index.html "The service is unavailable".  is in the corner of a blank page. 

Still this way at 5:55 pm eastern 7/21/2025. 

Could my configuration be the culprit? 

• Everything was smooth for weeks before that when I installed all the sensors with new batteries (that I hacked inside the housing). 
• The first time I saw hangs is when configuring the moisture alarms with high and low values. Not sure when I did this, sometime around July 11. Dragging the knob for these values would hang the UI, but direct numeric entry seemed OK. 
• But since then the service has been fine when using the Android app, but just cranky and crashing when using the web app, on either Chrome or Firefox. 
• We might have lost our Powerpuff plant, it was not ever showing alarms in the index of all sensors, even when dropping way below the 7%. All the rest of the alarms seemed to be working. I checked my settings, they are consistent with the working alarms, and interestingly, the graph shows orange below my threshold.  Screenshot 2025-07-21 173651.png  (but being in alarm state was not reflected in the sensor list that we were regularly checking on our phones)

In case it is my actual moisture alarms crashing the server or my use of the graphs on each moisture sensor, and nobody tells me otherwise, I will turn all the alarms off the next time the site is up. 

Jamie

[Dave M]

I agree that the site is back down; probably around 17:55 EDT like you said. There was an earlier post by the vendor saying a disk filled unexpectedly, but all is fixed now. Lol, maybe you are the one crashing everything after all! 

Such a great product. Didn't realize how much and often I use and check it! I've felt it today for sure.

[Zhiheng Cao]

No one sensor has been responsible for filling up the disk, the rate of filling has been pretty much constant. We manually move old data from SSD to HDD as SSD fills up on Google Compute Engine. This relies on periodic manual checking of empty space, which had a lapse. 

The 503 error simply means the server is too busy serving other requests.  The tag manager firmware has been historically designed to retry server request automatically no matter what, in a hindsight this was a design mistake that causes these issues following a server down time.  The longer the server downtime, the more tag managers there will be trying to call the server at the same time. 

The priority has been placed on ensuring no data is lost, especially for tag / sensors / probes with internal memory.  The tag managers also have internal buffers.  So if a tag manager gets a 503 error, it will keep the data in the buffer and retry automatically until it succeeds, at which point the server has stored the data in database. 

--
You received this message because you are subscribed to the Google Groups "Wireless Sensor Tags" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wireless-sensor-...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/wireless-sensor-tags/fe59cf95-9482-4fe7-86f4-e82384e5c1adn%40googlegroups.com.