Investigation: How many selected titles' cost-per-use return
- Pavlo Smahin
Owned by Pavlo Smahin
MODKBEKBJ-513 - Getting issue details... STATUS
Goal
Investigate the response time of GET /eholdings/packages/<packageId>/resources/costperuse endpoint.
Investigation
Endpoint's work steps:
- fetch title IDs from the holdings table;
- prepare requests to APIGEE
POST /uc/costperuse/titles. Limit up to 1000 title IDs per request. For publisher type = all - make 2 requests with parameter 'publisherPlatform' = true and 'publisherPlatform' = false - receive responses and generate mod-kb result
Response time comparison table
| Package | platform = publisher / nonPublisher | platform = all | |||
|---|---|---|---|---|---|
| ID | Amount of selected titles | Number of requests to APIGEE | Response time | Number of requests to APIGEE | Response time |
| 234-1310 | 100 | 1 | 1.5 - 2.4 | 2 | 1.5 - 2.6 |
| 184-1021 | 200 | 1 | 1.8 - 2.4 | 2 | 1.8 - 2.6 |
| 36-2758132 | 505 | 1 | 2.38 - 2.9 | 2 | 2.45 - 3.15 |
| 57-18766 | 998 | 1 | 2.86 - 3.5 | 2 | 3.05 - 5.6 |
| 36-4114 | 1430 | 2 | 2.98 - 3.7 | 4 | 4.42 - 6.0 |
| 372-2514919 | 3222 | 4 | 2.99 - 3.9 | 8 | 4.44 - 6.2 |
| 36-1510 | 6711 | 7 | 4.91 - 6.75 | 14 | 7.6 - 7.81 |
| 58-1322 | 22569 | 23 | 11.86 - 14.75 | 46 | > 20 |
Response time comparison chart
The biggest part of time in table is waiting responses from APIGEE. Calculation and sorting on mod-kb-ebsco side takes miliseconds.