Closed xdxxxdx closed 1 year ago
It would be interesting to make the scheduling strategy pluggable... (exponential back-off, simple time between requests,...)
It would be interesting to make the scheduling strategy pluggable... (exponential back-off, simple time between requests,...)
Currently when retry is enabled (soon it will be enabled by default), it comes with exponential random back-off by default. The retry http request executor expects a resilience4j config which allows us to easily add more options such as a simple time interval if this would be required in future use cases.
To be tested with Sprint 28
validated on ldes/ldi-orchestrator:1.0.0-SNAPSHOT
As lots of data source has a rate-limiting configured on their server.
To avoid receiving 429 responses all the time, I found it interesting to be able to configure the client's concurrency. For example, if we know the restrictions of a certain API we could say "do 1 page at a time" or "no more than X requests per 10 minutes"
https://github.com/Informatievlaanderen/VSDS-Linked-Data-Interactions/tree/main/ldi-orchestrator/ldio-connectors/ldio-ldes-client