Cap every scenario at 5 minutes (sample, not play-to-end)

Apple's bipbop-basic stream is genuinely 30 min long. Play-to-end across
all 4 scenarios was running ~85 min/leg — too slow for daily CI.

Switched all defaults to `fixed=300` (5 minutes of REAL playback):

  bipbop-adv:fixed=300
  bipbop-basic:fixed=300
  big-buck-bunny:fixed=300
  akamai-live:fixed=300

5 minutes is enough to verify decode + steady-state events. With the
10-second harvest cycle, that's ~30 harvest cycles per scenario.

Total per leg: ~25 min (20 min play + 5 min build/install/boot).
Total wall-clock for the matrix (4 legs in parallel): ~25 min.

The `end` mode is still in the script for ad-hoc deeper validation;
override DEFAULT_SCENARIOS or use the `nr_overrides` input if you
specifically need play-to-end on a one-off run.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: avinash-newrelic

Examples/iOS/NRSampleApp/README.md

@@ -22,14 +22,16 @@ The simulator stays focused so you can watch + hear playback. Artifacts land in
 
 ## What runs by default
 
-| Scenario | Stream | Mode | Duration |
-|---|---|---|---|
-| bipbop-adv | Apple HLS adaptive | play to natural end | ~10 min |
-| bipbop-basic | Apple HLS basic | play to natural end | ~30 min |
-| big-buck-bunny | Blender | play to natural end | ~10 min |
-| akamai-live | Akamai HLS live | 30 min cap | 30 min |
+Each scenario plays for **5 minutes of real playback** (a sample, not full playthrough), then is terminated. Total per leg ≈ 25 min (20 min play + 5 min build/install/boot).
+
+| Scenario | Stream | Cap |
+|---|---|---|
+| bipbop-adv | Apple HLS adaptive | 5 min |
+| bipbop-basic | Apple HLS basic | 5 min |
+| big-buck-bunny | Blender | 5 min |
+| akamai-live | Akamai HLS live | 5 min |
 
-VOD scenarios run until `didPlayToEnd` (capped at 60 min for safety). Live runs the full 30 min.
+5 minutes is enough to verify decode + steady-state events at the configured 10-second harvest cycle (≈30 harvest cycles per scenario). For deeper validation (memory leaks, hour-long stability), tune the cap up via the `nr_overrides` input or edit `DEFAULT_SCENARIOS` in `scripts/run-playback.sh`.
 
 ---
 

Examples/iOS/NRSampleApp/scripts/run-playback.sh

@@ -3,8 +3,8 @@
 # run-playback.sh — automation for the NRSampleApp catalog.
 #
 # Builds the app, boots a simulator, and plays every catalog item:
-#   - VOD: plays to natural didPlayToEnd (capped at VOD_SAFETY_CAP secs)
-#   - Live: plays for a configured wall-clock duration
+#   - Each scenario plays for 5 minutes of REAL playback (sample), then is
+#     terminated. ~25 min total per leg. See DEFAULT_SCENARIOS below.
 #
 # Real AVKit playback. Simulator window stays focused. Mid-playback
 # screenshot per scenario. Per-scenario log file copied to artifacts.
@@ -166,12 +166,16 @@ echo "    appToken=${NEW_RELIC_APP_TOKEN:+(set)}${NEW_RELIC_APP_TOKEN:-(not set)
 
 # ---- Scenarios -------------------------------------------------------------
 
-VOD_SAFETY_CAP=3600
+# Every scenario is capped at 5 minutes of REAL playback (a sample, not
+# play-to-end). Sampling is enough to verify decode + steady-state events;
+# play-to-end of Apple's 30-min bipbop-basic took ~85 min/leg — too slow
+# for daily CI.
+VOD_SAFETY_CAP=3600   # only used by `end` mode; defaults below use fixed=300
 DEFAULT_SCENARIOS=(
-  "bipbop-adv:end"
-  "bipbop-basic:end"
-  "big-buck-bunny:end"
-  "akamai-live:fixed=1800"
+  "bipbop-adv:fixed=300"
+  "bipbop-basic:fixed=300"
+  "big-buck-bunny:fixed=300"
+  "akamai-live:fixed=300"
 )
 
 if [ -n "${SCENARIOS:-}" ]; then