Ideology: Gatling vs Tank
Gatling and Tank are both open-source load testing tools, but they start from different places.
| Tank | Gatling (open-source edition) | |
|---|---|---|
| How tests are written | Recorded through the Tank Proxy, then cleaned up with filters and edited step by step in a web UI. | Written as code in Gatling's DSL (Java, Scala or Kotlin), often starting from its recorder. |
| Where tests live | In the controller's database, shared by everyone who uses the controller. | In source files, versioned with your code. |
| Where load comes from | Agents the controller starts on demand in AWS, in one or more regions. | The machine that runs the simulation. |
| Running and watching tests | Queued and controlled from the web UI or the REST API, with live users, TPS and failures per job and per agent in the Agent Tracker. | Run from the command line or a build tool, with an HTML report at the end. |
| Test data | Data files uploaded to the controller and split among agents. | Feeders defined in the simulation code. |
| Access control | Users, groups and ownership (see Roles and permissions). | None; anyone with the code can run it. |
Tank suits teams that want a shared, central place to record, keep and run load tests at scale without writing code. Gatling suits teams that want load tests as code alongside their application.