Kubernetes ConfigMap config store never completes when API server is unreachable

[InfoSec812]

When a Vert.x application making use of Kubernetes ConfigMaps is unable to connect to the Kubernetes API master it will hang indefinitely without failing... There needs to be a mechanism for it to fail after a specified amount of time.

package com.redhat.qcon

import io.vertx.config.ConfigRetrieverOptions
import io.vertx.config.ConfigStoreOptions;
import io.vertx.core.json.JsonObject
import io.vertx.reactivex.config.ConfigRetriever;
import io.vertx.reactivex.core.AbstractVerticle;
import io.vertx.core.Future;

public class DemonstrateConfigMapProblem extends AbstractVerticle {

    @Override
    void start(Future startFuture) {
        ConfigRetrieverOptions retrieverOptions = new ConfigRetrieverOptions()
        ConfigStoreOptions confOpts = new ConfigStoreOptions()
                .setType("configmap")
                .setConfig(new JsonObject()
                .put("name", "insult-config")
                .put("optional", true)
        );
        retrieverOptions.addStore(confOpts);
        ConfigRetriever.create(vertx, retrieverOptions)
                .rxGetConfig()
                .doOnError(startFuture.&fail)
                .subscribe({c -> startFuture.complete()})
    }
}

[cescoffier]

This issue needs to be handled in vertx service discovery and vertx config. Do you know a way to reproduce it?

[InfoSec812]

The snipped above should reproduce it quite handily. Just run it WITH the KUBERNETES_NAMESPACE env var set, but somewhere that there is no Kube master reachable.