How about increase opentelemetry plugin's plugin ordering?

[DingGGu]

Hello there,

With OpenTelemetry plugin with (header_type: aws) option can extract and use X-Amzn-Trace-Id Header.

- name: opentelemetry
  config:
    endpoint: 
    headers:
      api-key: 
    resource_attributes:
      service.name: ${CONFIG_OTEL_SERVICE_NAME}
    header_type: aws

This makes helpful for match between trace in APM tools and AWS LB access log.

However, My custom plugin has priority = 1000, So that plugin cannot get root's traceId, because opentelemtetry's plugin prioirty was 14. As you know, traceId was save kong.ctx.plugin.trace_id after opentelemetry plugin was executed.
In order to Plugin ordering document. Opentelemetry Plugin order was 14. I think this is too low.

1. What is your think about increasing opentelemetry plugin's priority?
   In last days, zipkin plugin is the most highest priority.

2. How about your think about decrase my plugin to negative prioirity?

[samugi]

I think it would be ideal to read headers and load the trace ID (and the rest of the tracing context) soon (high priority) and then report information to the observability backend late (low priority). One way to achieve that is by spliting the concept of headers propagation and reporting in separate plugins, so that the tracing context is loaded early and independently from the execution priority and configurations of tracing plugins (otel, zipkin, ...).