Skip to content

'Keda based Message Queue Trigger for NATS Jetstream' docs inaccuracies #174

@sunny-r

Description

@sunny-r

For the fields

TOPIC
RESPONSE_TOPIC
ERROR_TOPIC

It is documented (on github, and at the website) that they are "generally" of the form streamname.subjectname. Is it, or isn't it? What does 'generally' mean?

In my deployment, I could only get it working where those topics exactly matched only the subjects within a single stream (i.e. no separate stream for responses or errors):

fission mqt create \ --name flow-origin-trigger \ --function flow-origin \ --mqtype nats-jetstream \ --mqtkind keda \ --topic netflow-raw \ --resptopic netflow-labeled \ --errortopic netflow-error \ --contenttype='application/protobuf' \ --metadata stream=netflow-stream \ --metadata consumer=flow-origin-consumer \ --metadata natsServerMonitoringEndpoint=nats.nats.svc.cluster.local:8222 \ --metadata natsServer=nats://nats.nats.svc.cluster.local:4222 \ --metadata lagThreshold=30 \ --metadata account='$G'

Trying streamname.subjectname leads to 'subject not found' errors from NATS.

Also, at the website again: "Jetstream connector creates a push based subscriber to get the data. Make sure the consumer provided in mqt is of type pull. Also, if the consumer is not present connector will itself create the it."

This isn't clear at all, and in any case, I could only get a push-based consumer working.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions