Fix example config for Pulsar exporter, units are nanoseconds (#2185)

- setting partitions_auto_discovery_interval to 1 causes a storm of
  CommandPartitionedTopicMetadata requests to brokers. The broker doesn't log these
  requests and it's hard to detect the issue.
  The default value in the client is 1 minute which is 60*10^9.
  Set the example value to the default.

- setting batching_max_publish_delay to 10 will make batching uneffective
  since the unit is nanoseconds. The default value in the client is 10 milliseconds
  which is 10^7. Set the example value to the default.

internal/exporter/pulsarexporter/testdata/config.yaml

@@ -23,10 +23,12 @@ exporters:
       compression_level: default
       batch_builder_type: 1
       disable_batching: false
-      batching_max_publish_delay: 10
+      # unit is nanoseconds (10^-9), set to 10 milliseconds in nanoseconds
+      batching_max_publish_delay: 10000000
       batching_max_messages: 1000
       batching_max_size: 128000
-      partitions_auto_discovery_interval: 1
+      # unit is nanoseconds (10^-9), set to 1 minute in nanoseconds
+      partitions_auto_discovery_interval: 60000000000
 
 service:
   pipelines: