Performance issue with large request

NEST/Elasticsearch.Net version: 7.1.0. i also try 7.10.1 Elasticsearch version: 7.1.1 Description of the problem including expected versus actual behavior: A clear and concise description of what the bug is. We have built one API with .net and NEST. We had some issues with performance. So I have added timers that I log to check what is taking most of the time when we do a search Steps to reproduce:

private readonly ElasticClient elasticSearchClient;
        public ElasticSearchService(IHttpClientFactory httpClientFactory)
        {
            var settings = new ConnectionSettings(new Uri(appSettings.ElasticSearch.Host)).DefaultIndex(appSettings.ElasticSearch.PlaceIndexName);
            elasticSearchClient = new ElasticClient(settings);
        }
        private async Task<IList<Place>> SearchTextSuggestElasticAsync()
        {
            var time1 = DateTime.Now;
            ISearchResponse<Place> searchResponse = // some query;
            var requestTime = (DateTime.Now - time1).TotalMilliseconds;
            if ((requestTime - searchResponse.Took) > 300)
            {
                log.Warn($"the difference : {(requestTime - searchResponse.Took)} End request in {requestTime} took: {searchResponse.Took}");

            }
            var data = searchResponse.Documents.ToList();
            return data;
        }

It’s fine with a single request, but when I make about 100 request/s, requestTime with much larger than took from ElasticSearch I try to use HttpClient factory to do same number of request, it does not happen. So it does not network issue In above code, a lot of log was written the difference : 364.1491 End request in 529.1491 took :165 the difference : 351.0123 End request in 531.0123 took :180 the difference : 605.7492 End request in 747.7492 took :142 the difference : 587.7566 End request in 776.7566 took :189 the difference : 389.3551 End request in 492.3551 took :103 Expected behavior

[trongvodang]

Thanks, i will give it a try when i comeback to company

Vào Th 2, 25 thg 1, 2021 lúc 19:50 Steve Gordon [e-mail] đã viết:

@trongvodang https://github.com/trongvodang It’s linked in my previous comment - https://github.com/stevejgordon/ElasticsearchClientPerfTesting

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/elastic/elasticsearch-net/issues/5283#issuecomment-766793721, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIN2CVCRC4LEMNWIYQL3JVDS3VSITANCNFSM4WMTB36Q .

[stevejgordon]

@trongvodang Those counters look as I’d expect, no unusually high volume of connections. Thanks for running that. Please kill that cloud instance as I wouldn’t want those credentials to be misused. As part of our ongoing work for the next major release we will be focusing on performance and doing lots of profiling.