Provide a publisher mechanism to pass stored span data to a HTTP endpoint for post-processing about jaeger HOT 7 CLOSED

After further thought, as mentioned in the referenced discussion, having a kafka publisher would introduce dependencies in the server app. Having the http proxy may be a good way to go in the open source project, to decouple from the technology used for processing.

So as a first step, would it be possible to provide a mechanism to publish spans via HTTP if one or more endpoints are configured?

from jaeger.

publish in which format?

from jaeger.

Thrift would be good.

from jaeger.

Just to clarify - we would like to have an http endpoint in the collectors to save spans?

Or do we just want another storage sink that submit to some http endpoint?

from jaeger.

Providing a REST endpoint for reporting spans is also required, but covered by this issue: https://github.com/uber/jaeger/issues/123

This issue is to publish the span data to another http endpoint for post processing purposes.

from jaeger.

@objectiser would like to know more about the use case that requires this.

from jaeger.

This was based on an early discussion about how Jaeger could publish spans to a data pipeline via a REST API to avoid additional dependencies in the server - however I think more recent discussions about using something like Kafka to collect spans, and persist / post-process from there, have superceded such a requirement so will close.

from jaeger.