I’m trying to understand the recommended max message size in Hazelcast. The best documentation I can find is https://docs.hazelcast.com/hazelcast/5.0/capacity-planning where it says An optimum partition size is between 50MB – 100MB Does this mean messages should be smaller in size than that limit? Is there a way I can pass around messages that are 1Gb in size? Is this recommended?
Tag: publish-subscribe
Cold streams and publishers, is the Garbage Collector fully aware of its true death?
When instanceSubscriptorManager is unable to receive any more boolean values, the last thing to be performed will be a false as isActive, removing localConsumer from publisher. How is localConsumer “aware” that instanceSubscriptorManager will be unable to receive an isActive -> true (again) sometime in the future, especially since the class is self-referencing itself? In short, this class is hold by
Sending PubSub message manually in Dataflow
How can I send a PubSub message manually (that is to say, without using a PubsubIO) in Dataflow ? Importing (via Maven) google-cloud-dataflow-java-sdk-all 2.5.0 already imports a version of com.google.pubsub.v1 for which I was unable to find an easy way to send messages to a Pubsub topic (this version doesn’t, for instance, allow to manipulate Publisher instances, which is the