[doc] Add a pitch on the measurement of the performance of Sirius Web based applications

Signed-off-by: Stéphane Bégaudeau <stephane.begaudeau@obeo.fr>

Author: sbegaudeau

CHANGELOG.adoc

@@ -4,7 +4,7 @@
 
 === Pitches
 
-
+- Measure the performance of Sirius Web based applications
 
 
 === Architectural decision records

doc/iterations/2026.5/measure_the_performance_of_sirius_web_based_applications.adoc

@@ -0,0 +1,67 @@
+:author: Stéphane Bégaudeau
+:date: 2026-03-25
+:status: proposed
+:consulted: Pierre-Charles David
+:informed: Florian Rouëné
+:deciders: Stéphane Bégaudeau
+:issue: https://github.com/eclipse-sirius/sirius-web/issues/6326
+
+= (XS) Measure the performance of Sirius Web based applications
+
+== Problem
+
+We want to be able to measure the performance of Sirius Web based applications.
+
+
+== Key Result
+
+We should have a tool and some patterns to perform performance testing.
+It should allow us to measure the performance of the GraphQL HTTP API, the REST API and the WebSocket API simultaneously in one test if necessary.
+It should provide us with clear reports allowing us to track the performance of various use cases over time.
+We should be able to have keep some history of those performance reports.
+This history could be maintained manually in another document for now.
+
+The solution must be applicable to Sirius Web and downstream projects.
+It should be possible to reuse fragments of performance tests from Sirius Web in downstream applications.
+Those fragments should share the same philosophy as our reusable tests services such as query runners with the minimal amount of coupling necessary.
+
+
+=== Acceptance Criteria
+
+The performance test suites should be executable locally with a development instance.
+Such an instance should be able to have the frontend and backend running on two different ports (e.g. http://localhost:5173 and http://localhost:8080).
+They should also be able to run with a production instance where everything is packed into one executable.
+
+The performance test suites should also run during our continuous integration process.
+
+
+== Solution
+
+We will create a set of performance test suites based on reusable fragments that can be composed to write complex tests.
+It should also help initialize the state of the application with data that matches real usage patterns of Sirius Web.
+
+
+=== Scenario
+
+Sirius Web contributors should be able to write new performance tests and run existing ones.
+The manipulation of those performance tests should be straightforward and documented.
+
+
+=== Breadboarding
+
+No user interface needed
+
+
+=== Cutting backs
+
+We should be able to see a code coverage report of the application after the execution of the performance test suites to evaluate the parts of the application that have not be covered by the tests.
+
+
+== Rabbit holes
+
+We may identify some performance issues during our tests, we may not solve them but we will track them using Github issues.
+
+
+== No-gos
+
+None