6.2 KiB
Subscriber Search Service
- Architecture Notes
- Concerns
- Plan
- A Dedicated Service for Subscriber Search
- Expose search inputs backed with the existing database
- Enable dblink on the search master database
- Create new unlogged search results table
- Define the search result format
- Perform search using new search DSL
- Perform search using legacy segment ID
- Manage segments using the existing database
- Migrate to an updated schema
- Centralizing Subscriber Search
- Milestone 3: Add new search features
- A Dedicated Service for Subscriber Search
- Resources
A replacement for the current Sites Subscriber Search and Broadcast Segment Search implementations.
Architecture Notes
Added: [2019-08-06 Tue 10:34]
Broadcast sending manipulated queries before performing them to suit its needs
- Handle list exclusion in the search service? (CCPANEL-9557)
- Blocked emails
- De-duplication
- Lead view ids -> segment ids -> search service segments? (CCPANEL-9556)
Side-by-side comparison
- Use broadcast-segment to compare its results to the search api results
- Recipient-style representation or leads?
Define API (CCPANEL-9554)
- Include saved searches (segments)
Iterate release (CCPANEL-9555)
- Current UI
New UI with same capabilities of old UI
- Still writing old-style segments
Research getting off the leads table (aurora?)
- https://confluence.aweber.io/display/AR/Search+Service+Using+Existing+Databases
- Distinct from a materialized search-optimized db
- Deal with
subscriber_tags
table bloat in AppDB
Component Diagram
database Analytics {
database Ana as Ana01
database Ana as Ana02
database Ana as Ana03
}
database App
database "Results Cache" as ResultsCache
component "Search Service" as Service {
component Search
component Results
Search -- Analytics
Search -- App
Search --> ResultsCache
Results <-- ResultsCache
}
Concerns
Performance
Fitness to Purpose
This service will need to fulfill the needs of both end-user subscriber searches and segment emailing.
Plan
ITEM | Effort |
---|---|
Plan | 18d 0:00 |
Search Centralization | 18d 0:00 |
Expose search inputs backed with the existing database | 2d |
Enable dblink on the search master database | 2d |
Create new unlogged search results table | 2d |
Define the search result format | 1d |
Perform search using new search DSL | 5d |
Perform search using legacy segment ID | 3d |
Manage segments using the existing database | 3d |
Migrate to an updated schema | |
Migrate to new search service | |
Create new subscriber management React application | |
Update broadcast-segment to use new search service | |
Milestone 3: Add new search features |
A Dedicated Service for Subscriber Search
Create a dedicated, publicly exposed service for performing searches on subscribers using subscriber and analytics criteria. The goal of this project is to replace the current implementation from sites and in the broadcast segment service.
Expose search inputs backed with the existing database
- Include IDs required to build existing POST format
Enable dblink on the search master database
Create new unlogged search results table
Define the search result format
https://xd.adobe.com/view/ae8fb2b2-c039-4e88-8ade-ff2562a8c8cf-fbdc/screen/c03f09c7-187e-4a6f-8b8c-571d131daee1/ (ignore engagement column)
- name
- source
- status
- date added
- last updated
Perform search using new search DSL
Perform search using text comparisons
- Is / Is Not
- Contains / Does not contain
- Starts with / Does not start with
- Ends with / Does not end with
Perform search using numeric comparisons
Perform search using date ranges
Perform search using message and link IDs
Perform search using tag sets
Perform search using legacy segment ID
- Include flags for broadcast segment-required filtering?
Manage segments using the existing database
- Translate a DSL search into a legacy stored segment
Migrate to an updated schema
Centralizing Subscriber Search
Applications making use of subscriber search will be updated to use the new dedicated service, eliminating multiple search implementations.
Create new subscriber management React application
Update broadcast-segment to use new search service
Milestone 3: Add new search features
Resources
- PostgreSQL Backed Search (Rejected ACP)
- Search Proxy Service
[[https://confluence.aweber.io/display/~robink/SoT-ElasticSearch+Next+Steps][SoT - ElasticSearch Next Steps]]- Alternative Search Proposal
-
Search Service Using Existing Databases (Approved ACP)
- Search Service - Proof of Concept Findings (Benchmarks different approaches)
- Search Prototype
- Search DSL JSON Schema