SimpleSpanProcessor support in OTel

[brunobat]

Description

We need to support SimpleSpanProcessor in OTel, in order to allow lambda functions on cloud providers to report, in time their traces.
See #26250

The default processor accumulates spans and sends them in batches, therefore the it's called BatchSpanProcessor.
Lambda functions shut down quick and sometimes there is no time to send the traces out.
This feature will dispatch spans as they are created, therefore making sure the span goes out.

This applies to LogRecords as well.

Implementation ideas

This should be trivial after the OTel SPI access is available after the #26444 work is finished.
Should be a mater of documenting the procedure.

[geoand]

Is this still relevant?

[brunobat]

Very... Especially for lambda functions.

[brunobat]

FYI, This project can be used as a helper: https://github.com/brunobat/reproducers/tree/main/graalvm-aws-rest-26250

[brunobat]

Assigning to @alesj

[brunobat]

@arik-dig has this problem on tests that use openTelemetry. The test closes before the export finishes and throws an exception.
Really tight configs like these help but don't work all the time:

quarkus.otel.bsp.max.export.batch.size=1
quarkus.otel.bsp.schedule.delay=0.001S

[arik-dig]

when working with version 3.4.1 seeing the following warning:

19:57:32 WARNING traceId=, parentId=, spanId=, sampled= [io.op.sd.tr.ex.BatchSpanProcessor] (BatchSpanProcessor_WorkerThread-1) Exporter threw an Exception: java.lang.IllegalStateException: Client is closed
	at io.vertx.core.http.impl.HttpClientImpl.checkClosed(HttpClientImpl.java:696)
	at io.vertx.core.http.impl.HttpClientImpl.doRequest(HttpClientImpl.java:597)
	at io.vertx.core.http.impl.HttpClientImpl.request(HttpClientImpl.java:465)
	at io.vertx.grpc.client.impl.GrpcClientImpl.request(GrpcClientImpl.java:50)
	at io.quarkus.opentelemetry.runtime.exporter.otlp.VertxGrpcExporter.export(VertxGrpcExporter.java:91)
	at io.quarkus.opentelemetry.runtime.exporter.otlp.VertxGrpcExporter.export(VertxGrpcExporter.java:128)
	at io.opentelemetry.sdk.trace.export.BatchSpanProcessor$Worker.exportCurrentBatch(BatchSpanProcessor.java:326)
	at io.opentelemetry.sdk.trace.export.BatchSpanProcessor$Worker.flush(BatchSpanProcessor.java:271)
	at io.opentelemetry.sdk.trace.export.BatchSpanProcessor$Worker.run(BatchSpanProcessor.java:238)
	at java.base/java.lang.Thread.run(Thread.java:833)

as @brunobat mentioned when using the configs:

quarkus.otel.bsp.max.export.batch.size=1
quarkus.otel.bsp.schedule.delay=0.001S

it reduces the frequency of it happening, but its inconsistent.

[brunobat]

This has been implemented here: #43983