We have added Node.js support to allow users to write their own custom Node.js load agents in addition to the PHP support that has been available since last summer. We will be adding documentation and examples, but for those of you who want to experiment – go for it! Let us know any issues or if you have suggestions at info@redline13.com.Read More →
We did some upgrades this weekend on the Linux load agent so it will start faster. We also added a linux shutdown script that will inform Redline that the server is shutting down. This will help when people manually shut down servers.Read More →
With all of the large load tests that are running on RedLine now, we have made some big internal improvements that expand our scalability. This means that large users will not run into bottlenecks at the RedLine server level. Essentially, we have decreased our load on our database server with far fewer writes. We’ve replaced our Java SQS queue processor with one written in node.js. It screams now! So bring the load!Read More →
It is fun to watch the large scale tests that people are using RedLine for. Here is an example of a test with 10 m3.2xlarge servers each simulating 2000 users each hitting the applications 50 times from the AWS Ireland data center. So 1,000,000 hits on their application in a couple of minutes. Here is the graph of a successful test: As you can see the load was about 10,000 hits per second and theRead More →
RedLine is great for running long duration tests. In addition to being incredibly cheap, it is also easy. We give you two ways to set long duration tests. You can do the simple ping test by setting the # of Iterations parameter as seen in this example: The above test 100 users will randomly ping the URL between 0 and 20 seconds (20,000 milliseconds) – so about 5 requests per second. We have set theRead More →
Yes, you can do POST request types with RedLine. To do this, you will need to add your own Custom Test Script. You can find information and examples at https://www.redline13.com/blog/writing-a-custom-load-test/. If you download the first example file at the bottom, it will give you the basics of what you will need. The main thing you will change is to add an array of POST variables as the second argument to goToUrl. If you need any help, please letRead More →
When running a load test, make sure that your Load Agents are properly sized. We do an automatic “guesstimate” of the number of clients per load agent based on the size of the instance and previous experience. But since every test is different, you need to take a look at the graph at the very bottom of the page to see how the CPU utilization is doing. Here is an example of 25 load agents andRead More →
When writing a Custom test script, you need to add the ability to handle errors. Otherwise, the test script will cause that client to fail. This happens frequently because the request failed by receiving no response or timing out. Once it fails, that user test stops. If you want to continue after an error, you should wrap the goToUrl call in a try/catch. In the catch, add something like: catch (Exception $e) { // Record errorRead More →
When you chose your Load Agent deployment options, we do a real time call to AWS to find the current pricing of Spot Instances. We up this a bit to create an estimated bid price. However, if you are deploying a large number of spot instances of one type in one region, you can actually bid up the price of what is available. In this example, the current bid price of an m3.2XLarge is $0.25 perRead More →
We recently expanded support across all AWS regions. This means you can run load agents from any mix or match of regions. If you are testing an app only available in Japan, then test from there. If you are deploying a worldwide app, then have load agents deployed in each region.Read More →