[SPARK-37462][CORE] Avoid unnecessary calculating the number of outstanding fetch requests and RPCS

### What changes were proposed in this pull request?
Avoid unnecessary calculating the number of outstanding fetch requests and RPCS

### Why are the changes needed?
It is unnecessary to calculate the number of outstanding fetch requests and RPCS when the IdleStateEvent is not IDLE or the last request is not timeout.

### Does this PR introduce _any_ user-facing change?

No.
### How was this patch tested?
Exist unittests.

Closes #34711 from weixiuli/SPARK-37462.

Authored-by: weixiuli <weixiuli@jd.com>
Signed-off-by: Sean Owen <srowen@gmail.com>

Author: weixiuli

common/network-common/src/main/java/org/apache/spark/network/server/TransportChannelHandler.java

@@ -158,10 +158,10 @@ public void userEventTriggered(ChannelHandlerContext ctx, Object evt) throws Exc
       // To avoid a race between TransportClientFactory.createClient() and this code which could
       // result in an inactive client being returned, this needs to run in a synchronized block.
       synchronized (this) {
-        boolean hasInFlightRequests = responseHandler.numOutstandingRequests() > 0;
         boolean isActuallyOverdue =
           System.nanoTime() - responseHandler.getTimeOfLastRequestNs() > requestTimeoutNs;
         if (e.state() == IdleState.ALL_IDLE && isActuallyOverdue) {
+          boolean hasInFlightRequests = responseHandler.numOutstandingRequests() > 0;
           if (hasInFlightRequests) {
             String address = getRemoteAddress(ctx.channel());
             logger.error("Connection to {} has been quiet for {} ms while there are outstanding " +