A short intro about load testing with the remarkable tool that is Gatling.
% IANA WHOIS server
% for more information on IANA, visit http://www.iana.org
% This query returned 1 object
domain: vcz.fr
status: ACTIVE
hold: TIGHT
press: ➡️ TO CONTINUE
Oh no, here he goes about tests again 😭
credits: Unsplash
Ensuring, at different stages, that the service you are building, operates how you intended.
how: behaviors under extreme inputs, feature implementation, project architecture, dependencies, security, interactions between services, performance, error recoverability, etc.
Less deployment failures, development hell / churn / cost, quality issues, time to market, reputational damage.
Better deployment frequency, healthier teams, project durability / profitability.
credits: Mike Cohn
More expensive
Faster
Yes, but does it work well?
A well-developed application may still be subject to security, performance, architectural, resilience, cost, etc. issues.
You can of course follow logs, metrics, traces and KPIs after the fact but what if you shorten the cycle?
Today, we are interested in
But the other ones are cool too!
How each individual will use my service? Different user groups, profiles, personas have different impacts on our systems.
How a large population of individuals can impact our systems? What happens when they focus on a page and when they disperse? What effect did load on a page have on the others?
Request
VU
(Virtual User)
Scenario
+ think time
Inject a given number of VUs per unit of time.
No negative feedback: application will receive load at the same rate even if
down.
Suited for pushing queue-based systems.
Maintain a given number of active VU.
Negative feedback: application slows down, VUs complete slower.
Suited for pushing interactive systems.
Founded in 2012 in France 🐓
Open source and Enterprise
Heavy-weight of Load Testing
IDE plugins, protocol-agnostic, recorder, HTML reports, etc.
Synchronize the following commands:
👋