In my app, I’m using multiple imagery layers, and terrain data ( custom terrain data served from a third party server), the problem is one of the imagery layers (wmts ) has some tiles that are pending, when triggering rendering events ( e.g zooming in / out) the whole scene wouldn’t update until the pending tiles are loaded or timed-out ( server time out is 120 seconds). Is there a way to set a timeout policy for pending tiles in cesium ? I can see a TileDiscardPolicy but I don’t know how to utilize it.
Welcome to the community! Thank you for sharing some details about your issue. Do you have any idea why certain tiles are taking excessive amounts of time to load? I have not seen this behavior in the past.
TileDiscardPolicy seems like the correct tool to use here. Unfortunately, I do not have a sandcastle demo that showcases how to use this object. However, can share some documentation and a code snippet. According to the docs, TileDiscardPolicy should be called as such.
Thank you @sam.rothstein for your reply,
the issue in generating tiles comes from our backend, we haven’t encounter this when we were using other mapping libraries ( openlayers). Basically it struggles to generate low zoom levels requested by cesium imagery providers.
I noticed - if I understand correctly - that high level tiles must wait for parents -starting from level zero to be loaded at least until it gets rendered,
I was wondering if there is a way not to wait for these tiles to load, or to set up a timeout for them or skip them even (e.g. use a dummy tile instead).
Thank you for sharing an update with me! I do not think that our spec supports a way to set up a timeout for loading in tiles. However, your idea about using a dummy tile is intriguing. How exactly would you generate a tileset with these “dummy” tiles?
@sam.rothstein
I had to ditch the dummy image approach and go with setting a timeout to the loading module, so by editing Resource._Implementations.loadWithXhr function in our specific case and adding these lines to the request (line 2089) :
Thank you very much for sharing this update with me
I am happy to hear that you moved away from the dummy image approach. While that approach may have provided you with a temporary work-around, it was not a scalable solution.