Currently, a well-formed but incorrect URL supplied as --backend.sparqlEndpoint.url will successfully start but fail on every query. This is incorrect. There should be a health check that pings the sparql endpoint up front.
Use case
Makes operations easier and clearer.
Proposed solution
Just run a trivial query.
Additional information
Intriguing question is like, do we maintain liveness checks consistently? Seeing as SPARQL is not connection-based, what is the desired behavior of this container if the backend dies in the middle? Should it fail itself or continue to respond with 404s?
Feature Request
Packages to improve:
Description
Currently, a well-formed but incorrect URL supplied as
--backend.sparqlEndpoint.urlwill successfully start but fail on every query. This is incorrect. There should be a health check that pings the sparql endpoint up front.Use case
Makes operations easier and clearer.
Proposed solution
Just run a trivial query.
Additional information
Intriguing question is like, do we maintain liveness checks consistently? Seeing as SPARQL is not connection-based, what is the desired behavior of this container if the backend dies in the middle? Should it fail itself or continue to respond with 404s?