Hello. We are hosting a WMS with geoserver 2.27.2, using the Kartoza image in Docker.
However we have ran into issues with the rendering speeds with orthophoto in Qgis.
I am not a GIS engineer, just the software developer guy so I will try to explain it best I can.
When we render a heightmap for forest heights, we get relatively okay and good speeds of each request.
However when we try to use an orthophoto, the highest zoom level does not even complete the request before the timeout. And I am wondering why this is happening and what I could possibly do to improve it.
The issues does not seem to be the HDD’s, the %iowait is never high.
Checking the docker logs for the geoserver container, it seems that it accepts one request for each “view” in Qgis. So when I am at the maxlevel, it basically logs it as 1 request for the entire nation we are trying to display.
Each request is passed into 1 thread for 1 CPU. We have 20 cores, given the previous statement only 1 core is used to serve that big request, hence the timeout.
We also have 128GB available DDR4 memory, so that is not the bottle-neck either.
The forestheight is one file (.cog), where the orthophoto that is struggling is a layer-group consisting of approximately 20 cogs.
So how would I go about solving this, is it the datasources themselves that when constructed with gdal or whatever that needs to be different? Is there some setting in Geoserver I need to configure?
I did adjust the controlflow.properties to account for longer requests before the timeouts.
Sincerely