Batch Recommendation Deployment Plan
AWS
- Verify the AWS configuration in the staging environment.
Terraform @Luthfi
- Update terraform environment
PERSONALIZE_ROLE_ARN
USER_PERSONALIZATION_V2_SOLUTION_ARN
HH-Personalize Server @Luthfi
- Merge branch feat/batch-recommendation to main
- Release hh-personalize app
Exec Notebook Databricks @Luthfi
- Clear old data [on-progress]
- Load reservation data to s3 [on-progress 100%]
- Retry because of this error: ClickUp task #86cy1j39t
- Load restaurants data to s3 [on-progress]
- Load users data
- Verify data
- Execute hh-personalize-v2.ipynb
- Make sure env & aws account is correct
- Exec create s3 code first
- Update roles
- Create dataset group
- Create schema (interactions, items, users)
- Create schema dataset (interactions, items, users)
- Create Job import data
- Interactions [error][done]
- Items
- Users
- Create user personalization solution v2
- Save the output
- Run worker to roll out the recommendation [on progress 20%]
Check & Re-check @Luthfi
- Health check
https://personalize.hungryhub.com/health - GraphQL check
https://personalize.hungryhub.com/graphiql- Trigger deploy
- GetRecommendedForYou API
- BullMQ check
https://personalize.hungryhub.com/bullmq
HH-Pegasus
- Enable Homepage setting
Recommended For You@Luthfi - Live Review Production @Luthfi
- QA Review: @Bernadetta Kusumadewi @wiwik @Rina Apriani
- Monitor server @Luthfi
- Monitor worker @Luthfi
Post Release
- Ask QA & Thai Team to test
- If everything is good, enable it for all users
Jupyter Notebook Execution
Execute the Jupyter notebook to generate the batch recommendations.