I am looking for any tips/tricks/advice anyone may have on efficiently publishing WMTS services. I am experiencing processing times measured in hours to publish very small datasets. For instance I have a layer of roughly 100 manholes that cover an area less than 2 square miles that has been caching for 5 hours now and is only 40% through 11 of 12 zoom levels! I'm sure there's something off with my settings but frankly there isn't much documentation and given the time its taking trial and error isn't much of an option either.
Some thoughts/questions I've had:
Is there a way to tell the system what area cache (i.e the bounding box settings in a browser map view)?
Would the min/max display scales in the vector data settings help?
What about the buffersize factor in the vector data settings? I've left it blank thus far as I haven't noticed any clipping of features along the tile edges.
I've implemented scales in the styleset and that works for the display of the features but doesn't improve the publishing time.
What is the 'Depth' setting worked out to in this particular case? This will effect publish time.
The depth is calculated by the application when creating the 'my geoservice' based on the bbox and the scale parameter. It is not editable directly. To change the depth you would need to change the scale you are using. I larger scale value here will lead to a higher depth value. ie. scale of 50,000 will lead to a higher depth number than a scale of 1,000,000.
There isn't a way to set the specific area to cache, the buffersize vector setting and min/max display scales aren't helpful here. It really comes down to the scale parameter being set.
Thanks for the information, the Depth in this case ended up being 12 for a Scale of 25. Which is small I admit but my dataset covers a small area and some of the points are within a couple feet of each other so I need to be able to zoom in very close.
I'm assuming the scale/depth is directly related to the visibility of the features as you zoom in, so I started with a scale of 1000 (depth=6) then I tried 100 (depth=10) both of which cache in less than 1 hour but stop displaying as I zoom in. Unfortunately even at a scale/depth of 25/12 the features stop displaying before I get zoomed in close enough to discern them. Maybe I'm wrong in assuming the visibily is related to the scale/depth and don't need to be caching down this far in the first place?
