revert(load-generator): replace k6 with Locust (#3873)

* revert(load-generator): replace k6 with Locust

Reverts #3564. Restores the pre-k6 Locust-based load generator
(locustfile.py, requirements, Envoy /loadgen/ route, compose service
config) while keeping everything unrelated added since that commit.

This reverts commit 18b36c7.

Assisted-by: Claude Sonnet 5

* fix(load-generator): address Copilot review findings on Locust revert

- otel-collector: restore allowed_headers (content-type, traceparent,
  baggage) on the OTLP HTTP receiver's CORS config, otherwise the
  browser's baggage-bearing preflight requests are rejected and
  frontend-web telemetry disappears.
- locustfile.py: register the RoofBinoculars response listener before
  navigating instead of after, so a fast image response isn't missed.
- demo.flagd.json: restore loadGeneratorTraffic and loadGeneratorVUs,
  which the flagd-ui scheduler tests still assert exist.
- services.py: correct the frontend-web exclusion comment; browser
  traffic is enabled in every scope, it's just low-weighted relative
  to HTTP traffic, not disabled.
- README.md: point to the stable Locust docs instead of the 2.16
  archive, since the restored lockfile pins 2.44.4.

Assisted-by: Claude Sonnet 5

* Apply suggestion from @julianocosta89

* fix(flagd-ui): exclude non-off-resting flags from generic revert tests

loadGeneratorTraffic rests at "on" and loadGeneratorVUs rests at its own
default variant, not "off". With the fixed seed used by these tests, the
scheduler deterministically picked one of them, so the generic "reverts
to off" assertions never passed. Restrict those tests to off-resting
flags, matching the pattern already used elsewhere in this file.

Assisted-by: Claude Sonnet 5

* fix(load-generator): remove dead k6-era load-generator flags

loadGeneratorTraffic and loadGeneratorVUs were introduced by the k6
migration (#3564) and consumed only by its entrypoint.sh wrapper; the
restored Locust code never observed them. Remove both flags along
with the flagd-ui scheduler test cases and README section that
existed only to accommodate their non-off resting state, restoring
the pre-k6 flag set.

Assisted-by: Claude Sonnet 5

* fix(load-generator): use registered telemetry attributes

Replace unregistered span attributes (product.id, quantity, category,
item.count, flood.count) with their existing demo.* registry
equivalents, adding a new demo.request.flood.count attribute for the
one case with no existing match, and declare all of them for
service.load_generator per the telemetry-schema conventions.

Assisted-by: Claude Sonnet 5

* fix(load-generator): correct changelog and telemetry schema docs

- Restore the #3564 changelog bullets under 3.0.0, which was already
  released with them, and add a new Unreleased entry for this revert
  instead of rewriting released history.
- Fix the service.load_generator attribute group: its brief claimed
  "baggage attributes" but most of its members are plain span
  attributes, and it was missing user.id, which locustfile.py also
  emits as a span attribute.

Assisted-by: Claude Sonnet 5

* default flagd.json

---------

Co-authored-by: Juliano Costa <julianocosta89@outlook.com>
Co-authored-by: Juliano Costa <juliano.costa@datadoghq.com>

Author: mwimpelberg28

.env

@@ -100,8 +100,14 @@ IMAGE_PROVIDER_PORT=8081
 IMAGE_PROVIDER_DOCKERFILE=./src/image-provider/Dockerfile
 
 # Load Generator
-LOAD_GENERATOR_VUS=5
-K6_TARGET_URL=http://${FRONTEND_PROXY_ADDR}
+LOCUST_WEB_PORT=8089
+LOCUST_USERS=5
+LOCUST_HOST=http://${FRONTEND_PROXY_ADDR}
+LOCUST_WEB_HOST=load-generator
+LOCUST_AUTOSTART=true
+LOCUST_HTTP_USER_WEIGHT=9
+LOCUST_BROWSER_USER_WEIGHT=1
+LOCUST_HEADLESS=false
 LOAD_GENERATOR_DOCKERFILE=./src/load-generator/Dockerfile
 
 # Payment Service

.github/dependabot.yml

@@ -321,7 +321,7 @@ updates:
     patterns:
       - "*"
     multi-ecosystem-group: "src-chatbot"
-  - package-ecosystem: "gomod"  # zizmor: ignore[dependabot-cooldown] cooldown prevents multi-ecosystem batching
+  - package-ecosystem: "pip"  # zizmor: ignore[dependabot-cooldown] cooldown prevents multi-ecosystem batching
     directories:
       - "/src/load-generator/**"
     patterns:

CHANGELOG.md

@@ -7,6 +7,10 @@ the release.
 
 ## Unreleased
 
+* [load-generator] Revert #3564: replace k6 with the pre-k6 Locust-based load
+  generator, removing the now-unused `loadGeneratorTraffic` and
+  `loadGeneratorVUs` feature flags along with it
+  ([#3873](https://github.com/open-telemetry/opentelemetry-demo/pull/3873))
 * [react-native-app] Catch errors from `placeOrder` in the Cart screen and
   show an error toast so payment failures are visible to the user instead of
   being silently dropped

CONTRIBUTING.md

@@ -118,6 +118,7 @@ Once the images are built and containers are started, visit:
 - **Grafana**: [http://localhost:8080/grafana/](http://localhost:8080/grafana/)
 - **OpAMP**: [http://localhost:8080/opamp/](http://localhost:8080/opamp/)
 - **Feature Flags UI**: [http://localhost:8080/feature/](http://localhost:8080/feature/)
+- **Load Generator UI**: [http://localhost:8080/loadgen/](http://localhost:8080/loadgen/)
 
 The OpAMP UI shows the OpenTelemetry Collector as a managed agent. Click the
 collector instance ID to view its health status, version, attributes, and

Makefile

@@ -283,6 +283,7 @@ start:
 	@echo "Go to http://localhost:8080 for the demo UI."
 	@echo "Go to http://localhost:8080/jaeger/ui for the Jaeger UI."
 	@echo "Go to http://localhost:8080/grafana/ for the Grafana UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/feature/ to change feature flags."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 
@@ -294,6 +295,7 @@ start-minimal:
 	@echo "Go to http://localhost:8080 for the demo UI."
 	@echo "Go to http://localhost:8080/jaeger/ui for the Jaeger UI."
 	@echo "Go to http://localhost:8080/grafana/ for the Grafana UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/feature/ to change feature flags."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 
@@ -303,6 +305,7 @@ start-no-o11y:
 	@echo ""
 	@echo "OpenTelemetry Demo is running (no observability stack)."
 	@echo "Go to http://localhost:8080 for the demo UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/feature/ to change feature flags."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 
@@ -312,6 +315,7 @@ start-minimal-no-o11y:
 	@echo ""
 	@echo "OpenTelemetry Demo in minimal mode is running (no observability stack)."
 	@echo "Go to http://localhost:8080 for the demo UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/feature/ to change feature flags."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 
@@ -323,6 +327,7 @@ start-profiling:
 	@echo "Go to http://localhost:8080 for the demo UI."
 	@echo "Go to http://localhost:8080/jaeger/ui for the Jaeger UI."
 	@echo "Go to http://localhost:8080/grafana/ for the Grafana UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/profiles/ for the Firepit UI."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 
@@ -334,6 +339,7 @@ start-agentic:
 	@echo "Go to http://localhost:8080 for the demo UI."
 	@echo "Go to http://localhost:8080/jaeger/ui for the Jaeger UI."
 	@echo "Go to http://localhost:8080/grafana/ for the Grafana UI."
+	@echo "Go to http://localhost:8080/loadgen/ for the Load Generator UI."
 	@echo "Go to http://localhost:8080/feature/ to change feature flags."
 	@echo "Go to http://localhost:8080/telemetry/ for the Weaver generated telemetry documentation."
 	@echo "Go to http://localhost:8080/chatbot/ for interacting with demo application using an agent."

compose.yaml

@@ -330,6 +330,8 @@ services:
     environment:
       - FRONTEND_PORT
       - FRONTEND_HOST
+      - LOCUST_WEB_HOST
+      - LOCUST_WEB_PORT
       - GRAFANA_PORT
       - GRAFANA_HOST
       - JAEGER_UI_PORT
@@ -366,6 +368,8 @@ services:
     depends_on:
       frontend:
         condition: service_healthy
+      load-generator:
+        condition: service_started
       flagd-ui:
         condition: service_healthy
       telemetry-docs:
@@ -417,35 +421,37 @@ services:
     deploy:
       resources:
         limits:
-          memory: 512M
+          memory: 1500M
     restart: unless-stopped
+    ports:
+      - "${LOCUST_WEB_PORT}"
     environment:
-      - LOAD_GENERATOR_VUS
-      - K6_TARGET_URL
-      - OTEL_EXPORTER_OTLP_ENDPOINT=http://${OTEL_COLLECTOR_HOST}:${OTEL_COLLECTOR_PORT_HTTP}
-      - OTEL_EXPORTER_OTLP_PROTOCOL=http/protobuf
+      - LOCUST_WEB_PORT
+      - LOCUST_USERS
+      - LOCUST_HOST
+      - LOCUST_HEADLESS
+      - LOCUST_AUTOSTART
+      - LOCUST_BROWSER_TRAFFIC_ENABLED=true
+      - LOCUST_HTTP_USER_WEIGHT
+      - LOCUST_BROWSER_USER_WEIGHT
+      - OTEL_EXPORTER_OTLP_ENDPOINT
+      - OTEL_EXPORTER_OTLP_METRICS_TEMPORALITY_PREFERENCE
       - OTEL_RESOURCE_ATTRIBUTES=${OTEL_RESOURCE_ATTRIBUTES},service.criticality=low
       - OTEL_SERVICE_NAME=load-generator
+      - PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python
+      - LOCUST_WEB_HOST=0.0.0.0
       - FLAGD_HOST
       - FLAGD_PORT
       - FLAGD_OFREP_PORT
-      - K6_OTEL_EXPORTER_PROTOCOL=http/protobuf
-      - K6_OTEL_HTTP_EXPORTER_ENDPOINT=${OTEL_COLLECTOR_HOST}:${OTEL_COLLECTOR_PORT_HTTP}
-      - K6_OTEL_HTTP_EXPORTER_INSECURE=true
-      - K6_BROWSER_ENABLED=true
-      - K6_BROWSER_ARGS=no-sandbox,disable-dev-shm-usage
-      - K6_OTEL_METRIC_PREFIX=k6.
     healthcheck:
-      test: ["CMD-SHELL", "pgrep k6"]
-      start_period: 10s
+      test: ["CMD", "python3", "-c", "import socket; s=socket.create_connection(('localhost',${LOCUST_WEB_PORT}),2); s.close()"]
+      start_period: 30s
       interval: 5s
       timeout: 5s
-      retries: 5
+      retries: 20
     depends_on:
       frontend:
         condition: service_healthy
-      frontend-proxy:
-        condition: service_healthy
       flagd:
         condition: service_started
     logging: *logging

src/flagd-ui/README.md

@@ -30,9 +30,8 @@ Each interval the scheduler activates up to a configured number of distinct
 randomly picked flags. Every activation gets its own hold duration between the
 configured minimum and maximum, and its own offset, so that the whole activation
 fits inside the interval. When a hold expires that flag is set back to its
-resting variant, the flag's own configured default. That is `off` for most
-failure scenarios, but `on` for `loadGeneratorTraffic` and `5` for
-`loadGeneratorVUs`.
+resting variant, the flag's own configured default, which is `off` for every
+failure scenario.
 
 The following can be configured:
 

src/flagd-ui/test/flagd_ui/scheduler_test.exs

@@ -74,18 +74,6 @@ defmodule FlagdUi.SchedulerTest do
       assert {"cartFailure", ["10%", "100%", "25%", "50%", "75%", "90%"]} in flags
     end
 
-    test "offers loadGeneratorTraffic, whose resting state is on" do
-      flags = Storage |> GenServer.call(:read) |> Scheduler.schedulable_flags()
-
-      assert {"loadGeneratorTraffic", ["off"]} in flags
-    end
-
-    test "offers loadGeneratorVUs, whose resting state is its default variant" do
-      flags = Storage |> GenServer.call(:read) |> Scheduler.schedulable_flags()
-
-      assert {"loadGeneratorVUs", ["10", "25", "50"]} in flags
-    end
-
     test "tolerates a configuration without flags" do
       assert Scheduler.schedulable_flags(%{}) == []
       assert Scheduler.schedulable_flags(%{"flags" => nil}) == []
@@ -176,52 +164,13 @@ defmodule FlagdUi.SchedulerTest do
       assert Enum.any?(Scheduler.state(scheduler).history, &(&1.flag == flag))
     end
 
-    test "reverts loadGeneratorTraffic to on, not off" do
-      scheduler = start_scheduler(TrafficRevertScheduler)
-
-      config = immediate_config(%{flags: %{"loadGeneratorTraffic" => ["off"]}})
-
-      assert :ok = Scheduler.start_schedule(scheduler, config)
-
-      assert_receive {:scheduler_state, %{active: [%{flag: "loadGeneratorTraffic"}]}}, 2000
-      assert eventually(fn -> variant_of("loadGeneratorTraffic") == "off" end)
-
-      assert :ok = Scheduler.stop_schedule(scheduler)
-
-      assert eventually(fn -> variant_of("loadGeneratorTraffic") == "on" end)
-    end
-
-    test "reverts loadGeneratorVUs to its own default, not off" do
-      scheduler = start_scheduler(VUsRevertScheduler)
-
-      config = immediate_config(%{flags: %{"loadGeneratorVUs" => ["25"]}})
-
-      assert :ok = Scheduler.start_schedule(scheduler, config)
-
-      assert_receive {:scheduler_state, %{active: [%{flag: "loadGeneratorVUs", variant: "25"}]}},
-                     2000
-
-      assert :ok = Scheduler.stop_schedule(scheduler)
-
-      assert eventually(fn -> variant_of("loadGeneratorVUs") == "5" end)
-    end
-
     test "reverts every flag it activates when holds fill the whole interval" do
       scheduler = start_scheduler(BackToBackScheduler)
 
-      # Restrict to off-resting flags: loadGeneratorTraffic and loadGeneratorVUs
-      # rest elsewhere and are covered by their own revert tests above.
-      selection =
-        Storage
-        |> GenServer.call(:read)
-        |> Scheduler.schedulable_flags()
-        |> Enum.reject(fn {name, _} -> name in ["loadGeneratorTraffic", "loadGeneratorVUs"] end)
-        |> Map.new()
-
       # A hold as long as the interval makes each revert land exactly on the next
       # interval boundary, which previously left the earlier flag switched on.
       assert :ok =
-               Scheduler.start_schedule(scheduler, immediate_config(%{seed: 99, flags: selection}))
+               Scheduler.start_schedule(scheduler, immediate_config(%{seed: 99}))
 
       assert_receive {:scheduler_state, %{active: [%{flag: _}]}}, 2000
 
@@ -367,7 +316,11 @@ defmodule FlagdUi.SchedulerTest do
     test "reverts every concurrently held flag when stopped" do
       scheduler = start_scheduler(ConcurrentStopScheduler)
 
-      assert :ok = Scheduler.start_schedule(scheduler, held_config(%{concurrency: 4}))
+      assert :ok =
+               Scheduler.start_schedule(
+                 scheduler,
+                 held_config(%{concurrency: 4})
+               )
 
       assert_receive {:scheduler_state, %{active: [_, _, _, _] = active}}, 2000
 

src/flagd-ui/test/flagd_ui_web/live/scheduler_test.exs

@@ -23,7 +23,6 @@ defmodule FlagdUiWeb.SchedulerTest do
     assert html =~ "Scheduler stopped"
     assert html =~ "Idle"
     assert html =~ "adFailure"
-    assert html =~ "loadGeneratorVUs"
   end
 
   test "navigating from the dashboard live redirects", %{conn: conn} do

src/flagd/demo.flagd.json

@@ -93,24 +93,13 @@
         "on": 100
       }
     },
-    "loadGeneratorTraffic": {
-      "defaultVariant": "on",
-      "description": "Enable synthetic traffic from the load generator. Turn off to pause all load-generator scenarios.",
+    "loadGeneratorFloodHomepage": {
+      "defaultVariant": "off",
+      "description": "Flood the frontend with a large amount of requests.",
       "state": "ENABLED",
       "variants": {
         "off": 0,
-        "on": 1
-      }
-    },
-    "loadGeneratorVUs": {
-      "defaultVariant": "5",
-      "description": "Number of concurrent virtual users the load generator's HTTP scenario runs. Changes take effect on the wrapper's next poll, restarting k6.",
-      "state": "ENABLED",
-      "variants": {
-        "10": 10,
-        "25": 25,
-        "5": 5,
-        "50": 50
+        "on": 100
       }
     },
     "paymentFailure": {