How come my test burned more or less fuel than estimated?
When you’re designing your load test scenario, Loadster might display an estimate for how much Loadster Fuel the test will burn. This estimate is based on the number of bots, the type of bots, and the duration of each group in your scenario.
However, fuel consumption is based on the actual duration and number of bots started during the test. The actual test might not burn exactly the amount of fuel that was estimated.
The test ran longer than expected
Load tests sometimes run longer than the allotted time, especially if you have long scripts.
Bots normally repeat their script over and over, as long as there is time remaining in the test (we call each repetition an iteration). Even during the ramp-down phase, if a bot finishes and there is time remaining, the bot may start a new iteration.
Bots are allowed to finish their current iteration even the time runs out. If that iteration takes a while, it will delay the bot’s exit a while. Iterations can really take a while if your script has lots of steps, long wait times, or if your site is slow to respond.
The test stopped earlier than expected
Tests can also stop early in some situations.
Of course, if you terminate the test yourself, it will stop early. In this case it won’t burn as much fuel as was previously estimated.
Tests might also stop early if you have limited the iterations. If you limit the iterations, and the limit is reached while time still remains in the test, the bots will exit, and the test may finish early.