SQL Server - performance issue with concurrent queries

I recently ran into a problem with SQL Server issues. I am using node program to connect and run requests. As the number of requests / db increases, SQL Server performance degrades.

I used a basic test case where I get about 6 concurrent requests, which in turn causes six database requests in parallel. As a result, the evaluation time for each query is roughly multiplied by 6 times.

So basically, when I click on one request, it takes about 120ms, but in case 6 similar requests are in parallel, each request takes about 7500ms, which is troublesome with concurrent requests.

For the record, I am using SELECT

queries that should be non-blocking.

Any help or information would be deeply appreciated.

Update

I checked the table and it has no indexes. The request I tested is a simple SELECT

offer request WHERE

. I also checked sys.dm_tran_lock

for database locking information and found that it creates an IS lock on that table for every query request. SQL Server 2012 runs on an Azure instance. I'm not sure about the hardware spec, but it should be standard compliant.

I have been playing around with a slightly tricky combination of queries, it seems to be most striking when parallel queries are using the same table / view.

+3


source to share





All Articles