Oleksandr Porunov

@vojtechhabarta What do you think about it? Would it make sense to implement such a feature or do you see any problems if we relay on `null` value for optional...

I guess there are 2 internal issues: 1) `JtsPoint` isn't an instance of `JtsGeometry`. Thus, it is treated as simple point in `JtsGeoJSONWriter`. 2) Even if we add a special...

I don't want to block the release because I don't think we will work on fixing Z dimension soon. Later if we decide that Z dimension is needed we could...

With statefulset your hostname might be the same after re-deployment. In case you didn't properly close you JanusGraph instance and simply forced restart the pod then the instance id won't...

> Since JanusGraph needs to deserialize data which is a CPU-intensive step, the CQL thread pool (with a size close to the number of CPU cores) is still beneficial. That's...

> Let's say you have query.batch = True enabled and you have a query like g.V().hasId(within(1,2,3,4,...100k)).values(), there will be 100k queries to the storage backend. In this case, using the...

> That is not 100% accurate; it should be up to N times faster where N = min(pool size, number of CPU cores) I assumed that pool size is equal...

@li-boxuan I guess it makes sense to leave this CQL execution service but remove parallel backend executor service. That said, the purpose of this CQL execution service will change because...

> > That said, the purpose of this CQL execution service will change because it will be used for deserialization computation only and not for queries execution. > > I...

> this.session.executeAsync will submit the statement to the driver (which in turn sends the request to the Cassandra cluster). So the size of the thread pool won't limit the number...