Limit thread queue during init in ExchangeSource (#117273) (#117286)

ES|QL doesn't work well with 500 clusters or clusters with 500 nodes. 
The reason is that we enqueue three tasks to the thread pool queue,
which has a limit of 1000, during the initialization of the exchange for
each target (cluster or node). This simple PR reduces it to one task.
I'm considering using AsyncProcessor for these requests, but that will
be a follow-up issue for later.

Author: dnhatn

x-pack/plugin/esql/compute/src/main/java/org/elasticsearch/compute/operator/exchange/ExchangeSourceHandler.java

@@ -220,20 +220,21 @@ void onSinkComplete() {
      * @see ExchangeSinkHandler#fetchPageAsync(boolean, ActionListener)
      */
     public void addRemoteSink(RemoteSink remoteSink, int instances) {
-        for (int i = 0; i < instances; i++) {
-            var fetcher = new RemoteSinkFetcher(remoteSink);
-            fetchExecutor.execute(new AbstractRunnable() {
-                @Override
-                public void onFailure(Exception e) {
-                    fetcher.onSinkFailed(e);
-                }
+        fetchExecutor.execute(new AbstractRunnable() {
+            @Override
+            public void onFailure(Exception e) {
+                failure.unwrapAndCollect(e);
+                buffer.waitForReading().listener().onResponse(null); // resume the Driver if it is being blocked on reading
+            }
 
-                @Override
-                protected void doRun() {
+            @Override
+            protected void doRun() {
+                for (int i = 0; i < instances; i++) {
+                    var fetcher = new RemoteSinkFetcher(remoteSink);
                     fetcher.fetchPage();
                 }
-            });
-        }
+            }
+        });
     }
 
     /**