Skip to content

Latest commit

 

History

History
78 lines (47 loc) · 2.72 KB

JMeter.md

File metadata and controls

78 lines (47 loc) · 2.72 KB

JMeter Tests

Set-Up

Install & Set Up JMeter

  • Download JMeter (should be >= version 5.4.1).

  • Add the latest DocumentDbDriver .jar to /lib folder of your JMeter installation.

Set Up a Test Cluster

If targeting a cluster that has not been used for JMeter testing before,
you will need to insert the relevant test data beforehand. Use mongoimport to insert the data from /testData, using the name of each file as the collection name and “jmeter” as the database.

A potential enhancement could be to automate this.

Run Tests

Run Test Plan with DocumentDb Driver

GUI

  1. Open the DocumentDb_Test_Plan.jmx file in JMeter.

  2. Start an SSH tunnel for the target cluster.

  3. Click on the DocumentDb Test Plan element. Change user variables CONNECTION_STRING, USERNAME, and PASSWORD depending on SSH tunnel setup and target cluster.

  4. Run the test plan.

Command-line

  1. Start an SSH tunnel for target cluster.

  2. Run the DocumentDb_Test_Plan.jmx file. Pass connection string (-JconnectionString), username(-Jusername), and password(-Jpassword) as parameters. Connection string and username may be omitted as they have default values but password is required. Example:

    ./jmeter.sh -n -t DocumentDb_Test_Plan.jmx -Jpassword=<password>

Run Test Plan with Other JDBC Driver

  • It may be useful to run the tests against another data source to confirm expected functionality. The user variables CONNECTION_STRING, USERNAME, and PASSWORD can be changed to be used with another driver.

  • Dependencies of the other driver may need to be setup and/or added to the JMeter /lib folder. Refer to JMeter documentation and documentation of the specific database vendor for more information.

Add Tests

Add Test Data

Add Test Cases

  • To add a query, you need to add a new row in Test_Plan.csv.

  • Populate the query column with the query to be executed, the result column with the expected result set, and the test_name column with a descriptive name for the query.

  • To keep results uniform across different data sources for easier comparisons:

    • Columns are ordered explicitly instead of using SELECT * so column order is deterministic.

    • Rows are ordered explicitly with ORDER BY so row order is deterministic.

    • Calculated fields are named explicitly with AS so they always use same name.