How to disable or modify Thrift receiver in Jaeger v2 (Port 14268 active by default) ?

[gokulvootla]

Hello Maintainers 👋,

Thank you for your hard work on Jaeger and the recent v2 developments — it's exciting to see Jaeger align more closely with the OpenTelemetry ecosystem.

I’m working with Jaeger v2 and trying to disable or change the Thrift HTTP receiver port (default: 14268). Even after modifying the ConfigMap to change the port (or exclude the Thrift receiver entirely), the Jaeger Collector still appears to start and listen on port 14268, as seen in the pod logs.
I’ve confirmed the updated config is mounted correctly inside the container, and I’m passing it using the --config-file argument. Despite this, the Thrift HTTP receiver is still active on its default port.

Is there a way to disable or override the Thrift HTTP receiver in Jaeger v2?
Or is it currently hardcoded and always initialized regardless of configuration?

Any guidance on how to fully disable or control that receiver would be helpful, especially in a setup where we want to use only OTLP as the input and avoid exposing unused ports.

Thanks in advance for your support, and happy to test or contribute where possible!

[yurishkuro]

please share your config

[gokulvootla]

extensions:
  healthcheckv2:
   endpoint: 0.0.0.0:13133
  remote_sampling:
   file:
     path: /etc/sampling/samplingstrategies.json
     reload_interval: 0s
   http:
     endpoint: 0.0.0.0:5778
   grpc:
     endpoint: 0.0.0.0:5779
  jaeger_storage:
   backends:
     collector_storage:
       elasticsearch:
         auth:
           basic:
             username: ""
             password: ""
         server_urls: <host>:9200
         tls:
           cert_file:
           key_file:
           ca_file: 
           min_version: "1.2"
           max_version: "1.3"
           reload_interval: 3m
exporters:
  jaeger_storage_exporter:
   trace_storage: collector_storage
processors:
  batch: {}
receivers:
  jaeger:
   protocols:
     grpc:
       endpoint: 0.0.0.0:14250
       tls:
         cert_file: 
         key_file: 
         client_ca_file: 
         client_ca_file_reload: true
         reload_interval: 3m
         cipher_suites:
          
         min_version: "1.2"
         max_version: "1.3"
       keepalive:
         server_parameters:
           max_connection_age: "0"
           max_connection_age_grace: "0"
  otlp:
   protocols:
     grpc:
       endpoint: 0.0.0.0:4317
       tls:
         cert_file: 
         key_file: 
         client_ca_file:
         client_ca_file_reload: true
         reload_interval: 3m
         cipher_suites:
           
         min_version: "1.2"
         max_version: "1.3"
       keepalive:
         server_parameters:
           max_connection_age: "0"
           max_connection_age_grace: "0"
     http:
       endpoint: 0.0.0.0:4318
       idle_timeout: "0s"
       read_header_timeout: "2s"
       read_timeout: "0s"
       tls:
         cert_file: 
         key_file:
         client_ca_file:
         client_ca_file_reload: true
         reload_interval: 3m
         cipher_suites:
          
         min_version: "1.2"
         max_version: "1.3"
  zipkin:
   endpoint: 0.0.0.0:9411
   tls:
     cert_file: 
     key_file: 
     client_ca_file: 
     client_ca_file_reload: true
     reload_interval: 3m
     cipher_suites:
      
     min_version: "1.2"
     max_version: "1.3"
service:
  extensions:
   - jaeger_storage
   - healthcheckv2
   - remote_sampling
  pipelines:
   traces:
     exporters:
       - jaeger_storage_exporter
     receivers:
       - otlp
       - jaeger
       - zipkin
     processors:
       - batch
  telemetry:
   logs:
     level: debug
   metrics:
     level: detailed

[gokulvootla]

Tried to change thrift_http protocol from the default 14268 to a different port (e.g. 14279). Our updated configuration looks like this:

receivers:
  jaeger:
    protocols:
      thrift_http:
        endpoint: 0.0.0.0:14279
        idle_timeout: "0s"
        read_header_timeout: "2s"
        read_timeout: "0s"

However, after deploying this change, the pod logs still show the collector starting the HTTP server on port 14268:

info    jaegerreceiver@v0.127.0/trace_receiver.go:336    Starting HTTP server for Jaeger Thrift    
{"resource": {"service.name": "jaeger"}, "otelcol.component.id": "jaeger", "otelcol.component.kind": "receiver", "otelcol.signal": "traces", "endpoint": "localhost:14268"}

We’ve verified the following:

• The updated configuration is correctly mounted in the pod.

• The endpoint is clearly set to 0.0.0.0:14279.

• The Jaeger receiver is listed under the traces pipeline in the service section.

• The collector was restarted after applying the changes.

service:
  pipelines:
    traces:
      receivers:
        - otlp
        - jaeger
        - zipkin
      processors:
        - batch
      exporters:
        - jaeger_storage_exporter

[yurishkuro]

diff --git a/cmd/jaeger/config.yaml b/cmd/jaeger/config.yaml
index f94e9e26..c716191e 100644
--- a/cmd/jaeger/config.yaml
+++ b/cmd/jaeger/config.yaml
@@ -72,6 +72,7 @@ receivers:
       thrift_binary:
       thrift_compact:
       thrift_http:
+        endpoint: 0.0.0.0:54321

   zipkin:

$ go run ./cmd/jaeger --config cmd/jaeger/config.yaml

Log: 2025-07-18T11:16:39.735-0400 info jaegerreceiver@v0.129.0/trace_receiver.go:336 Starting HTTP server for Jaeger Thrift {"resource": {"service.instance.id": "0d265978-52a1-4e79-b558-ccaaf2009ed8", "service.name": "jaeger", "service.version": ""}, "otelcol.component.id": "jaeger", "otelcol.component.kind": "receiver", "otelcol.signal": "traces", "endpoint": "0.0.0.0:54321"}

Looks to me it's your set up that is at fault.

[gokulvootla]

Thanks @yurishkuro

I tried to change the port for thrift to 14098. Yes, It happened but strange is that still 14268 port is beeing used by the thrift first.

{"message":"2025-07-21T09:03:12.800Z\tinfo\tjaegerreceiver@v0.127.0/trace_receiver.go:336\tStarting HTTP server for Jaeger Thrift\t{"resource": {"service.name": "jaeger"}, "otelcol.component.id": "jaeger", "otelcol.component.kind": "receiver", "otelcol.signal": "traces", "endpoint": "localhost:14268"}","metadata":{"container_name":"collector","namespace":"test","pod_name":"new"},"service_id":"collector","severity":"info","timestamp":"2025-07-21T09:03:12.801+00:00","version":"1.2.0"}

{"message":"2025-07-21T09:03:11.707Z\tinfo\tjaegerreceiver@v0.127.0/trace_receiver.go:336\tStarting HTTP server for Jaeger Thrift\t{"resource": {}, "otelcol.component.id": "jaeger", "otelcol.component.kind": "receiver", "otelcol.signal": "traces", "endpoint": "0.0.0.0:14098"}","metadata":{"container_name":"collector","namespace":"test","pod_name":"new-4"},"service_id":"collector","severity":"info","timestamp":"2025-07-21T09:03:11.707+00:00","version":"1.2.0"}

Changes: I made change to the configmap

thrift_http:
endpoint: 0.0.0.0:14098
idle_timeout: "0s"
read_header_timeout: "2s"
read_timeout: "0s"