Provide a smoother experience on low bandwidth networks

[sbouchet]

Which functionality do you think we should update/improve?

from https://plugins.jetbrains.com/plugin/12030-openshift-toolkit-by-red-hat/edit/reviews#review=101851

It just hung there when expand namespace(project). Even worse, it froze my whole IDE(PyCharm). So please test this plugin in low speed network condition (e.g 50KB/s). Thanks.

Why is this needed?

Provide a smoother experience on low bandwidth networks.

[adietish]

@sbouchet: the namespace is fetched via intellij-kubernetes while the child items are retrieved via odo. We should of course prevent freezes but not sure about being able to reduce the requests by odo.

Looking into the code I spot odo and helm being the culprits. Here's what's executed when you expand the namespace:

intellij-openshift-connector/src/main/java/org/jboss/tools/intellij/openshift/tree/application/ApplicationsTreeStructure.java

Lines 128 to 144 in 76d50fd

	private Object[] createNamespaceChildren(@NotNull NamespaceNode namespaceNode) {
	OdoFacade odo = root.getOdo().getNow(null);
	if (odo == null) {
	return new MessageNode[]{new MessageNode<>(root, namespaceNode, "Could not get project children")};
	}
	if (!odo.namespaceExists(namespaceNode.getName())) {
	return new MessageNode[]{new ChangeActiveProjectLinkNode(root, namespaceNode)};
	}
	List<Object> nodes = new ArrayList<>();
	nodes.addAll(getComponents(namespaceNode, odo));
	nodes.addAll(getServices(namespaceNode, odo));
	nodes.addAll(getHelmReleases(namespaceNode));
	return nodes.toArray();
	}

1. odo.namespaceExists is executed via the kubernetes client and shouldn't take much time.
2. Then the odo binary is used to getComponents, getServices and the helm binary to getHelmReleases. This takes the bulk of bandwidth/time. There isn't much we can do to improve this. The only thing I see is to display helm charts in a separate view.

Thoughts?