Hi Gerard,

Query performance has never been this bad. Currently the lag is over 6 HOURS.. and rising

My previous question stands.. What is the plan because we do not cope.

how about each hosts their own? This would provide some relief.

Below, I attached a query to https://databus.dbpedia.org/repo/sparql to query the latest download urls for the Wikidata-dbpedia extraction: https://databus.dbpedia.org/dbpedia/wikidata

Here is the yasgui link: https://tinyurl.com/yy768vh3

We have a virtuoso docker image that takes the query, downloads the files and fills a local sparql endpoint:

1. Download the Dockerfile https://github.com/dbpedia/dev.dbpedia.org/raw/master/pics/Dockerfile.dockerfile
   
2. Build docker build -t databus-dump-triplestore .
3. Load any Databus ?file query:

   docker run -p 8890:8890 databus-dump-triplestore $(cat file-with-query.sparql)

Doing it this way would ease some load and the docker updates each week and can be cronjobbed. 

Note that this is for the Wikidata-DBpedia extraction: http://svn.aksw.org/papers/2015/ISWC_Wikidata2DBpedia/public.pdf

Databus is an open platform, so as soon as Wikidata/WMF or somebody else publishes the original wikidata dumps there, you can use the docker to decentralise hosting.

All the best,

Sebastian

QUERY: