To install and run the jmeter test:
- Download JMeter from here http://tweedo.com/mirror/apache//jmeter/binaries/apache-jmeter-2.11.zip
- Unzip it to a directory of your choice (will be called ${jmeter-home} in this guide
- Make binaries executable ( chmod +x ${jmeter-home}/bin/jmeter )
- Run JMeter ( cd ${jmeter-home} => ./bin/jmeter )
- Open SemaGrow Stack JMeter Workspace ( semagrow-stack-jmeter.jmx )
Situation until now :
- You have a JMeter workspace that's running two SPARQL queries agains dbpedia ( SELECT * WHERE { ?s ?p ?o } LIMIT 1 (and 2) )
- You have a JMeter workspace that's simulating one ( = 1) user running two queries once ( = 1 time ) => results in two ( = 2 ) requests to dbpedia.
- Run this test : ctrl+R
Checkout output of Listeners :
- In JMeter open "Thread Group" => "HTTP Request"
- Click on the nodes with the "meter"-icon (all these listeners will have different output)
- note documentation of listeners: http://jmeter.apache.org/usermanual/listeners.html
Adapting settings :
-
Thread Group 1.1. Number of Threads (users) : simulates the number of users 1.2. Loop Count (should correspond to the number of queries (lines) in SparqlQueries.csv
-
HTTP Request Defaults (these settings will be taken for all HTTP Requests) 1.1. Server Name or IP : the endpoint's domain 1.2. Path : the path to the endpoint
-
HTTP Request 1.1. Parameters : key=query, value=${query} [ = query from SparqlQueries.csv ]
-
SPARQL Queries from CSV 1.1. Definition on the source file for the SPARQL Queries (one query per line)