fix: prevent HTTP connection reuse in BlobClient error recovery

util/headerreader/blob_client.go

@@ -214,11 +214,11 @@ func (b *BlobClient) GetBlobsBySlot(ctx context.Context, slot uint64, versionedH
 		blobs, err = b.getBlobs(ctx, slot, versionedHashes)
 	}
 	if err != nil {
-		// Creates a new http client to avoid reusing the same transport layer connection in the next request.
-		// This strategy can be useful if there is a network load balancer in front of the beacon chain server.
-		// So supposing that the error is due to a malfunctioning beacon chain node, by creating a new http client
-		// we can potentially connect to a different, and healthy, beacon chain node in the next request.
-		b.httpClient.Store(&http.Client{})
+		// Create a new HTTP client with a dedicated transport that disables connection reuse.
+		// With the default client (nil Transport), Go reuses the global DefaultTransport and its connection pool,
+		// which may keep using the same problematic backend connection. Disabling keep-alives forces a fresh TCP
+		// connection on the next request, increasing the chance of hitting a healthy backend behind a load balancer.
+		b.httpClient.Store(&http.Client{Transport: &http.Transport{DisableKeepAlives: true}})
 
 		b.useLegacyEndpoint = !b.useLegacyEndpoint