Sensor data doesn't transfer to NS site

[olaluukkainen]

• Uploader Version Number: 0.7.1
• Your Android Phone Model Name (e.g. Samsung J1 Ace): Samsung S5 and S7
• Android Version Number (e.g. use 4.4 for 4.4.4): 8.0.0 (S7) 6.0.1 (S5)
• Network Connection at the time the issue occurred (Wi-Fi, Mobile): Both mobile and Wi-Fi
• (if relevant) Nightscout website version (e.g. 0.8.4 Funnel cake, master / dev branch): 14.0.6

All of a sudden both of our uploader phones stopped sending sensor data. Treatments, micro boluses, IOB, pump info etc. is visible in NS site. No errors in uploader phones. Sensor data is shown correctly in uploader phones.

No error codes. Migration to Atlas is done. Papertrail has this message repeating: "previous bucket not found, not calculating delta"

Severity Score

4

[Pogman]

Could be an issue with your NS database being near full. Go into your NS admin page and clear out some of the old data.

[olaluukkainen]

Don't know what the problem was but both uploaders started to work after re-installing the software. Sensor data still doesn't show in NS from the blackout period.

[olaluukkainen]

So this is still an issue. The problem started again today on both phones.

[pazaan]

@olaluukkainen Is this still occurring? If so, I'd need to set up some time so that we can remote debug this together. I no longer have 600 Series sensors, so I'm unable to test myself.

[funkstille]

Today both of my phones had the same issue. No data is transferred to nightscout anymore.

Switched the phone, same problem
Updated nightscout on heroku to latest, same problem
Downgraded uploader from 0.73 to 0.71, same problem

I switched from mongo to atlas round about 2 months ago, so that should not be the root cause as well. Database size is 108 MB ~ 21% of the free Space. Any suggestions?

It looks like only cgm data is not transferred. Manual blood sugar measurements are uploaded.

Forwarding data to xDrip+ and uploading via xDrip is possible.

[Pogman]

Make sure your using the latest version of Nightscout currently 14.0.7

[funkstille]

Yes, upgraded to nightscout 14.0.7 after the problem occurred. Did not solve the problem. Workaround is forwarding to xDrip and uploading with xDrip

[Pogman]

There were issues pre-14.0.7 that would not handle cgm backfills and other cgm issues. Have you tried again just running with an uploader reinstall done after 14.0.7 ?

[funkstille]

I did a upgrade to 14.0.7 and since that had no effect I have done a reinstall with downgrade to 0.71. Did not work. But since it is pretty late in Germany I might need a few hours of sleep. I will have look tomorrow morning, the backup phone has still 0.73 installed.

But everything worked fine until tonight. Even with nightscout lower then 14.0.7 and uploader on 0.73.

[funkstille]

I purged the uploader on both phones and reinstalled version 0.73.
Still no cgm data upload. Other data like ‘usb device failure’ is uploaded.
I also recognised that the app crashes from time to time.
At the moment I am using drip+ as uploader.

Should I provide any logs or other data? Screensharing?

[Pogman]

0.7.2 and 0.7.3 are have crashes for some people and that should get sorted soon.

I don't know what could be wrong with your cgm data specially if your seeing the issue with 0.7.1 too.

Can you go into the debug settings and enable "extended errors" Go back to the log screen and you should see items like this:

The nightscout info tells us what happened after a successful upload. The E:1/0/1 part is for cgm entries and there should be something similar for you after each 5 min poll.

Also look back through the log for this line:

This will show the exact time difference and version that the uploader is seeing. Can you check yours?

[funkstille]

Thanks you for your reply. This morning at about 11 o’clock it started to work again. Until that time I haven’t had the logs switched to extended errors. So sadly there is nothing to share from before. Right now the time difference between my nightscout instance 14.0.7 changes every now and then but is always less then 60 seconds. If the problem appears again I will log the information!

[Pogman]

The extended errors option just toggles what is visible and will always show the last few days whenever you turn it on.

The NS time difference will always vary but it should be a small amount and well under 5 minutes. So that looks good and not causing your issue.

Do you run multiple uploaders? This should not be a problem and even running simultaneously but maybe there's a clue.