Are there any statistics about how often manually dirtying to initiate a re-rendering of tiles occurs per day? If it would occur just once in a while (1 time per minute?) would it make sense to set a higher priority to render these tiles earlier than normal tiles in the render queue? asked 18 Dec '16, 17:18  katpatuka |
2 Answers:
I cannot answer the first question, but to answer second ("would it make sense to set a higher priority to render these tiles earlier than normal tiles in the render queue?"): From my point of view NO, dirty tiles should be rendered with lower priority than regular rendering requests, since many tiles for map viewing are rendered on the fly. This means that someone, who is browsing an area that didn't have tiles already rendered, needs them ASAP when viewing the area. It would have a very bad impact on user experience if users are sitting in front of a white map, waiting for their on-the-fly tiles to be rendered, while the server is busy with rendering tiles that where tagged dirty - a dirty map tile is better than no tile at all. answered 19 Dec '16, 12:31  jot |
Hi katpatuka, the first question could be answered by the OSM statistics builder Pascal Neis, he can be reached here http://neis-one.org/ He might be able to answer the second question. But for starters you should not ask 2 questions in one question. answered 18 Dec '16, 22:00  Hendrikklaas |
I do not think Pascal Neis keeps statistics on the number of webserver requests. This data is not available in the data dumps he uses in his typical analysis.