Skip to content

HAR analysis: Launchpad → Event view

Source: launchpage -> control room event stram.har — Launchpad → select event view → click control room → Event stream.

Flow: User is on the launchpad, switches to Event view, then opens a control room to see the event stream (events + assets panels).

Duplicated API calls (observed in first ~5s)

API Count Notes
GET /controlrooms/{id} 5× Same control room fetched 5 times (worse than Map/Assets)
GET /controlrooms/{id}/collaborators 4× Collaborators loaded 4 times
POST /events/search (initial) 4× Events search called 4× with same params before polling starts
POST /cycles/summary?size=250 4× Two pairs of identical calls (likely 2 components × 2 subscriptions)
POST /tasks/{id}/search 3× Task count fetched 3 times
GET /controlrooms/{id}/assets?size=500 3× Same initial assets load 3 times
GET /controlrooms/{id}/assets?size=500&page=1 3× Same pagination page 3 times
GET /controlrooms/{id}/places 2× Places loaded twice
GET /controlrooms/{id}/events/summary 2× Event summary (for insights panel) twice
GET /controlrooms?size=100 2× Full control rooms list twice
GET /controlrooms/{id}/config 2× Config twice

Asset polling duplicates

  • assets?recordedAfter=... often fires 2× in quick succession with the same recordedAfter timestamp (e.g. 23:07:17 + 23:07:18, 23:07:21 + 23:07:22, 23:07:27 + 23:07:28).

Comparison with Map/Assets

API Map/Assets Event view
getControlRoom 4× 5×
getCollaborators 3× 4×
events/search (initial) 2× 4×
cycles/summary 2× 4×
tasks/search 2× 3×
assets (initial + page 1) 1+1 3+3

Event view has more duplication than Map/Assets, likely because more components (event panel, event summary, assets panel, insights) each trigger their own loads.

Improvement opportunities

  1. Same as Map/Assets: deduplicate getControlRoom, getCollaborators, getControlRooms, getConfig.
  2. events/search — 4× initial is severe; consolidate to single load, shared via store.
  3. cycles/summary — 4× (2 pairs); find both call sites and dedupe.
  4. tasks/search — 3×; single source for task count.
  5. assets — 3× initial + 3× page 1; route through store, load once.
  6. Asset polling — debounce/dedupe when same recordedAfter fires twice.

Estimated impact

  • Before: ~45+ API calls in first 5s (initial load only).
  • After fixes: ~15–20 calls — ~55% reduction.