Note: the information below is for Demand plans only


You may know by now that we charge according to how many node hours you use. Here's an article to explain what exactly a node hour is. But how do you know how many node hours you'll need?

It's not easy to answer this question without further information because there are a few things that can significantly increase or decrease your cost. Let's go over some of these factors.

The Number of Users

How many concurrent users do you need?
Obviously, the number of users you decide to run will influence your cost. However, what's not so obvious is that the number of requests that you want to hit your application with does not necessarily correlate to the number of concurrent users you need to run. 

As an example, if you need 1000 requests within 5 minutes, that does not necessarily mean 1000 concurrent users. While it could be 1000 users sending one request within the 5 minutes, it could also be 1 user sending 1000 requests. You can vary the speed at which a user sends a request, allowing you a bit more leeway.

We recommend that you consider how many concurrent users you actually need, and use the minimum amount of users to generate the load you require. You need to think about how important it is to have a certain number of concurrent connections vs. just having the requests hit your server. There will be instances, such as when testing to see whether your server can handle a certain amount of user sessions, where concurrent users matter, of course. But if you're just looking at the load sent to your application, try running with fewer users that send more requests. This will reduce how many nodes you need, and how much you'll need to pay as well.

The Load Test Tool

What load test tool do you want to use?

Just like any application, each load test tool requires resources (CPU, memory) to run. Because each tool executes load tests slightly differently, they also have different requirements. On one of our nodes, here's how many users each of the tools we support can execute on average:
JMeter - 1,000 users
Gatling - 1,000 users
Selenium - 5 users
Flood Element - 50 users

As a result of this disparity, protocol-level tools like JMeter and Gatling will always be more cost-efficient in terms of the number of nodes you need to run than browser-level tools like Selenium and Flood Element.

The Duration

How long do you want to run your test?

Simply knowing how many users you want to run is not enough; how long you want to run them will impact your bottom line. One thing to keep in mind when you're planning your tests is that our minimum charging interval is 12 node minutes. This means that if you run a five-minute test and run nothing else on that node, you'll still be charged 12 node minutes or 0.2 node hours. With this in mind, estimate the total duration of your tests for the month in factors of 12 minutes (ex: 12, 24, 36, 48, 60 minutes).

Click here for a bit more information on optimising your usage of Flood and keeping your costs low.

The Location

From which regions do you need to run tests from?

Another factor is whether you need to run tests from a particular geographic region. This is relevant because you will need to start up a grid in every region you want to inject load from, even if all the users would normally have been able to be executed from a single node.

For example, while you can run 1000 users from one node, if you want to inject load from US East, Singapore, Ireland and Sydney, you'll need to start four nodes with 250 users each. You will also be charged for each node, meaning that you'll require 48 node minutes (12 node minutes x 4 nodes) for a single 12-minute test.

How much will this cost?

So how do node hours correlate with cost? Let's go over a few common scenarios to see what you can get for some of our more popular plans. You can, of course, refine this further; however here's an idea of what you can get.

20 node hour plan for $149 a month (for month-to-month contracts) / $99 (for yearly contracts)

If you're only going to run a handful of one-off tests, then the standard On Demand plan with 20 node hours will probably suit your needs.

  • 20,000 users with JMeter/Gatling for one hour
  • 1,000 users with JMeter/Gatling for 20 hours
  • 100 users with Selenium for one hour
  • 5 users with Selenium for 20 hours
  • 1,000 users with Flood Element for one hour
  • 50 users with Flood Element for 20 hours

50 node hour plan for $299 a month (for month-to-month contracts) / $240 (for yearly contracts)

  • 50,000 users with JMeter/Gatling for one hour
  • 1,000 users with JMeter/Gatling for 50 hours
  • 250 users with Selenium for one hour
  • 5 users with Selenium for 50 hours
  • 1,000 users with Flood Element for one hour
  • 50 users with Flood Element for 50 hours

1000 node hour plan for $2,250 a month (for month-to-month contracts) / $2,000 (for yearly contracts)

  • 1,000,000 users with JMeter/Gatling for one hour
  • 1,000 users with JMeter/Gatling for 1000 hours
  • 5,000 users with Selenium for one hour
  • 5 users with Selenium for 250 hours
  • 50,000 users with Flood Element for one hour
  • 50 users with Flood Element for 250 hours

For more specific requirements and questions about other plans or Hosted plans, please don't hesitate to contact us via chat or at support@flood.io . We'd be happy to help you out in more detail.

Did this answer your question?