Plan on Search implementation was presented
- Question on whether it's more efficient to send IDs as Kafka messages with Indexer calling back to Inventory/SRS to get actual records vs sending the data on the Kafka message queue. Already discussed - will evaluate approach as work takes place
- Discussion on 3 deliverables in Phase 1. Deliverable 1 doesn't impact the FOLIO UI. That comes with Deliverable 2. Deliverable 3 allows the project to support ElasticSearch as part of CI/CD pipeline and reference environments
- Phase 1 vs Phase 3 - TBD how many search use cases can/should be handled by Phase 1 capabilities. But it's expected that for sure would be accurate hit counts.
- Need to recognize that changing search behavior and capabilities while welcomed, needs to be coordinated with SIGs and POs to make sure that expectations continue to be met.
- Would be very beneficial to have a Product Owner for search to navigate all of the requirements, etc.
- Question: If Iris is the target release for Phase 1, what are the resources recommendations? Roughly - 3 developers, mileage may vary based on specific requirements.
- Possible way to implement Phase 1 without getting mired in all the various requirements is to expose ES as a new "Type" - peer to Query Search"
- Conclusion:
- Are TC OK with the proposed Phases and deliverables?
- Should we consider a non-ElasticSearch option aside from PostgreSQL Full Text Search (which has already been evaluated). Some have experience with Solar Cloud which has it's own challenges. Much of this has already been discussed and we feel ES is the best option to explore.
- Still concern with the way Kafka is planned to be used
- YES
- Is there anything else needed from TC?
- No - can be proposed to Cap Planning to evaluate resources and fit for Iris (R1)