-
Notifications
You must be signed in to change notification settings - Fork 2.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
jaeger-agent can directly connect to kafka ? #1499
Comments
I don't think the memory queue introduces any delay unless the collectors are under-provisioned. Nobody asked for this feature so far, so we don't have any plans around it. |
I may not be very clear about what I said. Jaeger-agent can connect directly to kafka, which can save the cost of deployment (no collector is needed), because now jaeger-agent is connected to the load balancer (haproxy) by domain name, and then the load balancer is responsible for distributing the request to jaeger-collector (currently two machines). If the jaeger-agent can connect directly to kafka, then for us to save at least 3 machines and actually a small time delay |
@wuyupengwoaini did you consider sending data directly to the collector instead of the agent? Client libraries should support HTTP sender which is able to send data directly to the collector. |
@pavolloffay languages with asynchronous capabilities can theoretically support direct delivery to jaeger-collector. A small part of our company's core system uses asynchronous support very poor php language, so we can only use jaeger-agent to reduce performance loss. |
Correct, the OTEL components will allow you to configure any combinations of the receivers/exporters. I am closing this as done. |
We are preparing to use jaeger's tracing data for real-time analysis and offline analysis. So we use jaeger-collector to store the data in kafka, then jaeger-agent is still connected to jaeger-collecor.
But I found that the link is too long, and there is a memory queue inside each jaeger component, so the analysis will have a relatively large delay. But if jaeger-agent can connect directly to kafka, it can not only improve real-time performance, but also eliminate the need to deploy jaeger-collector separately.
So, does Jaeger have plans to support jaeger-agent to connect directly to kafka?
The text was updated successfully, but these errors were encountered: