# On campus meeting 8/4/2021

## Agenda

- [x] feedback PR edit tour modal
- [x] performance testing bespreken
- [x] datastructuur task
- [x] Metrics service design
- [ ] General UI redesign (colors, fonts)
- [ ] asset library & asset modal design
- [ ] feedback during image MC task

## Notes

### Performance testing

- Main bottleneck on server = I/O
- Unacceptable service degradation:
-

#### First performance test

- Separate load testing of tour requests, task requests and media requests (images & 3D models)
- Measure avg response time and max requests per minute

#### Second performance test

- Mix of tour, task and media requests
- sample from random distribution weighed on expected request frequency
- Measure max requests/min before unacceptable service degradation
- (optional if time)

#### Third performance test

- Realistic traces from application
- Measure max users/hour before unacceptable service degradation

### Detective Task (AR)

- 3 chances
- after each try, if wrong, the user is told if they are close or not (hot or cold)
- There is a "goal" zone, if the user taps within the zone, the answer is correct
- The user must confirm their placement after tapping
- Points are awarded based on how many tries
- Feedback:
- if correct or wrong after 3 tries: regular feedback modal
- Else: on screen text: hot(ter) or cold(er)

#### Datastructure

- Coordinate of point
- Question string
- Media Image Url
- DetectiveTaskType (indicates if it is a question is image based)

### Metrics Service

#### First metrics to collect:

- Starting a POI
- Closing a POI
- Starting a tour
- Closing a tour

#### Metric Service Architecture

- InfluxDB
- Metrics from application are sent to single HTTP endpoint and are committed to InfluxDB

#### Metrics fields:

- tour_id
- user_id
- timestamp
- type (TOUR_START, TOUR_STOP, POI_START, POI_STOP)
- type specific fields

#### Metrics Retrieval

- Try to use Grafana
- otherwise: compute own analytics + chart.js

### asset library & asset modal design