No data in WCS responses

I’ve been using the Sentinel Hub WCS service (via eolearn/senhub py) frequently in the last few days. Some time in the last 24 hours, all my requests return 0s only.

Example request:

Investigating further: At one point this morning, my EO patches were coming back empty after executing the usual tasks which fill them with Sentinel 2 data. Now they are just coming back with 0s instead of any real data.

I am trying to resolve same problem, it occured about a week ago and still lasts, even though I have 400 000 Processing Units, I am not able to download the data through eolearn / sentinel-hub py.
Please, share you progress if possible.
Thanks in advance :slight_smile:

I don’t know if this is a sporadic problem. I checked my cache - I have real data downloaded from 2am this morning, but nothing has worked since I started a new process at ca. 9 am this morning.

So something clearly went wrong this morning, but I made a mistake in the test script which was causing the no data. Now I try my main program again, it is working fine.

I don’t know if related or relevant - while testing, I tried this page:
When I put my instance ID in (which has a TRUE_COLOR) layer, the WCS service picture would go blank. Playing with the MAXCC, and switching the input to L1C would sometimes make it show up but there is some strange behaviour I didn’t figure out…

Hi @barrett,

this specific call is asking for data just outside the border of the orbit, see the image, your AOI is in the centre of the image. So it seems to me that there are actually no data there…

@dansuske, can you be more specific with your problem? Ideally sending an example of a request, like what barret did? So that we can analyse your problem?

My mistake - I put the wrong crs in the request…
Looks like a much more interesting part of the world than intended! :slight_smile:
Thank you for taking a look.

Hm, weird. We did not notice anything strange in the morning and there was no deploy or similar.

I’m getting quite a lot of 500 responses now - my code handles it fine, but maybe this was happening earlier too?

Looking the stats there were 32.667 successful requests in the last hour and 130 5xx errors, so 0.4%.
This is not out of ordinary as there are things that can go wrong, occasionally, e.g. virtual machine dying, object storage errors, etc.

I am working with eo-learn script, I am using a geojson to choose the data I want to download, it already downloaded the data more then a week ago, so the data is present. I get only warning but the script won´t download the data. However, i get these warning messages:

C:\Users\daniel.suske\AppData\Local\Continuum\anaconda3\envs\EO_learn_2\lib\site-packages\sentinelhub\download\ SHUserWarning: In order to achieve faster download performance please set configuration parameters ‘sh_client_id’ and ‘sh_client_secret’
warnings.warn(message, category=SHUserWarning)
C:\Users\daniel.suske\AppData\Local\Continuum\anaconda3\envs\EO_learn_2\lib\site-packages\ RuntimeWarning: invalid value encountered in true_divide

The warning is there only to suggest you should try using Process API rather than WCS.
You will find information on how to get Authentication token here:

That being said, this warning does probably not relate to the fact that you are not getting any data. Some other issue must be present in your code.