Segment Skill

This skill helps you understand, debug, and navigate segment-related code across the Datalayer platform. Segments allow users to define custom filters and rule groups to partition analytics data (sessions, users, people, companies, web conversions, CRM objects) into dedicated BigQuery tables for reporting and analysis.

Invocation

When the user asks about segments, use this knowledge base to:

1. Identify which files, lambdas, and query builders are relevant
2. Explain the end-to-end segment creation, rebuild, and reporting flow
3. Help debug segment-related issues by pointing to the right logs, tables, caches, and code
4. Explain the rule group structure, condition types, and query generation

Repositories

What Are Segments?

Segments are user-defined filter rules that group analytics data into subsets. Each segment:

• Targets a specific object type (session, user, people, company, web conversion, or CRM objects)
• Contains one or more rule groups with conditions (property, metric, relationship, event-based)
• Generates a dedicated BigQuery table (segment_{segmentId}) containing matching object IDs
• Can be applied to reports to filter dashboard and analytics data
• Supports incremental updates via cronjob queries

Every account has a default "All Data" segment (non-custom, no rules, always active).

Segment Objects (What You Can Segment)

ListenLayer Objects

CRM Objects (Salesforce Integration)

Condition Focus Types

Each rule group uses one condition focus type:

Segment Field Types

RuleGroups Data Structure

ruleGroups: [
  {
    mainOperator: 'include' | 'exclude',      // Include/exclude matching records
    conjunction: 'and' | 'or',                // Combine conditions within rule
    segmentObject: 'session' | 'user' | ...,  // Target object type
    conditionFocus: 'property' | 'metric' | ..., // Condition focus type
    embeddedSegmentId: 'uuid-or-null',         // (Optional) ID of another segment embedded into this rule group
    dateRangeFilter: {
      type: 'lifetime-of-the-object' | 'previous-period' | 'specific-date-range',
      numberOfDaysPrevious: 30,
      startDate: '2024-01-01',
      endDate: '2024-01-31'
    },
    // Conditions (g0, g1, g2, ...)
    g0: {
      conjunction: 'or',
      type: 'dimension' | 'metric' | 'custom' | ...,
      key: 'propertyName',
      subKey: 'subPropertyName',
      subKeyLv2: 'deepPropertyName',
      condition: 'equals' | 'contains' | ...,
      value: 'filterValue',
      property: 'eventProperty',
      id: 'conversionId'
    },
    // ... g1, g2, etc.
  }
]

API Endpoints (datalayerapp-v2)

File: server/routes/segments.js

Segment CRUD

Embedded Segments

Segment Options & Preview

Segment Reports

Other

Server-Side Architecture

Controller

File: server/controllers/segments.js

Thin middleware layer that extracts request data and delegates to services. Each function receives req.body or req.params and calls the corresponding service function with accountTimeZone from the active account.

Service

File: server/services/segments.js

Core business logic (key functions):

Segment Sync Service

File: server/services/segmentSync.js

Handles cascade rebuilds when an embedded segment is updated:

Segment Testing Service

File: server/services/segmentTesting.js

Function handleSegmentTesting: Tests segment queries for validity across all object types and condition focus combinations. Validates paths against event schema. Returns query validation results and error reporting.

Model

File: server/models/segments.js

Segments Table Operations

SegmentReport Table Operations

Database Tables

Segments Table

Migration: server/migrations/20240517005932createtable_segments.js

SegmentReport Table

Migration: server/migrations/20240530040609tcreatetable_segmentReport.js

BigQuery Query Builders

Directory: server/models/query-builders/common/segment/

Main Query Builder (index.js)

Sub-Builders

Query Generation Flow

getSegmentData({ segment, accountId, accountTimeZone? })
  |
  v
Auto-fetch accountTimeZone from AccountModel if not provided
  |
  v
For each ruleGroup:
  getRuleQuery({ group, conditionFocus, segmentObject })
    |
    +-- PROPERTY/METRIC --> getObjectDataByPropertyMetric()
    |     +-- SESSION --> sessionObject.js
    |     +-- USER --> userObject.js
    |     +-- PEOPLE --> peopleObject.js
    |     +-- COMPANY --> companyObject.js
    |     +-- WEB_CONVERSION --> webConversionObject.js
    |     +-- CRM_* --> crmObject.js
    |
    +-- WEB_CONVERSION_EVENT/ECOMMERCE_EVENT/SALES_CONVERSION_EVENT
    |     --> getObjectDataByOptimizedView()
    |
    +-- RELATIONSHIP --> getObjectDataByRelationship()
    |
    +-- OTHER_EVENT --> getObjectDataByOtherEvent()
    |
    +-- If embeddedSegmentId present:
          +-- WITH conditions: AND objectId IN (SELECT objectId FROM segment_{embeddedId})
          +-- WITHOUT conditions: SELECT DISTINCT objectId FROM segment_{embeddedId}
  |
  v
Combine rules with AND (INTERSECT DISTINCT) or OR (UNION)
  |
  v
Return: { createTableQuery, cronjobQuery }

Lambda Functions (datalayerapp-v2-function-code)

1. HandleSegmentBigQuery

• Path: lambda-functions/HandleSegmentBigQuery/
• Handler: index.js
• Trigger: SQS (HANDLESEGMENTBIG_QUERY)
• Purpose: Processes segment cronjob queries for all accounts using SQS self-queuing for level-by-level execution
• Operating Modes:

- Initial invocation: Receives accountId and timeZone from SQS, fetches segments, computes dependency levels, starts processing level 0 - Continuation invocation: Receives serialized state (levels, pendingJobs, failedSegments, checkCount), checks if previous level's BigQuery jobs are complete, advances to next level or re-queues

• Constants: CHECKDELAYSECONDS = 180 (3 min between status checks), MAXCHECKCOUNT = 10 (30 min max wait per level)
• What it does:

1. Receives SQS message with accountId and timeZone 2. Queries PostgreSQL for active, non-blocked segments via SegmentModel.findByAccountId() 3. Uses getProcessingOrder() to topologically sort segments into dependency levels 4. Serializes levels (keeping only: id, name, query, ruleGroups) for SQS state passing 5. Calls processNextLevel() for level 0: - filterSegmentsByDependencies() — excludes segments whose embedded dependencies failed - submitLevelJobs() — submits all BigQuery batch jobs in parallel via Promise.allSettled. For each segment: parses ruleGroups, replaces date placeholders ({{startDate}}, {{endDate}}) via replaceDateRangeQuery(), sanitizes table name, calls queryBQBatch() - Last level: Submits jobs and returns without waiting (no dependents need the result) - Non-last levels: Self-queues SQS message with 180s delay containing pendingJobs, levels, failedSegments, checkCount 6. On continuation invocation: checkPendingJobs() checks BigQuery job statuses via checkBigQueryJobStatus(): - All done → advance to next level via processNextLevel() - Still running and checkCount < MAXCHECKCOUNT → re-queue with incremented checkCount - checkCount >= MAXCHECKCOUNT → treat remaining as failed, advance to next level 7. After all levels complete: processSSTriggers() reads server-side trigger rules from S3 (server-side-trigger/{accountId}.json). If People/Company/Sales Conversion Event targets exist and hour < 12 in account timezone, fetches schedule triggers via getAllScheduleSSTrigger() and sends SQS to schedule report queue (chunked in batches of 100, 900s delay)

• Filters (findByAccountId): isBlocked = false, queryStatus = 'done', query IS NOT NULL, totalBytesBilled < 10 (GB), platformAccountStatus NOT IN ['Canceled', 'Dormant (permanent)', 'Dormant (temporary)']
• Key files: models/segments.js, lib/bigquery.js, s3/readFileStaticJson.js, helpers.js, utils/dependencyGraph.js, utils/index.js
• Key functions:

- submitLevelJobs(segments, accountId, timeZone) — Parse ruleGroups, replace date placeholders, submit BigQuery batch jobs in parallel. Returns { success, segmentId, bigQueryJobId } per segment - checkPendingJobs(pendingJobs) — Check BigQuery job statuses. Returns { allDone, failedSegmentIds } - filterSegmentsByDependencies(segments, failedSegments) — Exclude segments with failed embedded dependencies - processNextLevel(params) — Orchestrate level processing: filter → submit → self-queue or return - processSSTriggers(accountId, timeZone) — Post-processing: send scheduled report triggers - sendScheduleTriggerSavedReport(triggers, accountId, timeZone, targetObject) — Chunk and send SQS messages for schedule triggers

Dependency Graph Utility (HandleSegmentBigQuery)

• Path: lambda-functions/HandleSegmentBigQuery/utils/dependencyGraph.js
• Purpose: Topological sorting and cycle detection for segment processing order
• Exports:

- extractEmbeddedSegmentIds(segment) — Extract embedded IDs from a segment's ruleGroups (handles string or object). Returns empty array on parse error - buildDependencyGraph(segments) — Build nodes/edges graph from segment array. Only includes edges where both segments exist in current dataset - hasCycle(graph) — DFS-based cycle detection using recursion stack - topologicalSort(segments) — Kahn's algorithm using reverse dependency map and in-degree counting. Returns segments grouped by processing level. Falls back to single level on cycle detection - getProcessingOrder(segments) — Entry point. Fast path (single level) if no segments have embeddings, otherwise runs full topological sort

BigQuery Integration (HandleSegmentBigQuery)

• Path: lambda-functions/HandleSegmentBigQuery/lib/bigquery.js
• Exports:

- queryBQBatch({ query, accountId, type, tableName }) — Creates BigQuery batch job with useLegacySql: false, priority: 'BATCH', labels for accountid/type/tablename. Returns job object or undefined on error - checkBigQueryJobStatus(jobId) — Retrieves job metadata, returns { jobId, status }. Status values: PENDING, RUNNING, DONE, ERROR, NOT_FOUND

Utilities (HandleSegmentBigQuery)

• Path: lambda-functions/HandleSegmentBigQuery/utils/index.js
• Exports:

- parseJSON(jsonString) — Safe JSON parse, returns null on failure - getDateFormat(date, timeZone, format) — dayjs date formatting with timezone - splitIntoChunks(arr, limit=100) — Chunk array for batching SQS messages - replaceDateRangeQuery(query, ruleGroups) — Replace {{startDateN}} / {{endDateN}} placeholders using dateRangeFilter.numberOfDaysPrevious from the Nth ruleGroup

2. HandleCheckBigQueryJob

• Path: lambda-functions/HandleCheckBigQueryJob/
• Handler: index.mjs
• Trigger: SQS (AWSQUEUECHECKBIGQUERY_JOB)
• Purpose: Monitors BigQuery job completion for segment creation/rebuild
• Retry limiting: Skips processing if ApproximateReceiveCount > 1 (prevents duplicate processing)
• Input validation: validateInput() checks required fields (jobId, accountId, action), validates action type against BQJOBACTION enum, parses jobInputData as array
• What it does:

1. Receives SQS with jobId, accountId, action, jobInputData 2. Validates input, finds job in DB, skips if job status is not in-progress 3. Checks all BigQuery job statuses via checkBigQueryJobStatus() for each bigQueryJobId in payload 4. If not all jobs completed: - Tracks executionCount (increments per check, 15s per execution) - If under max time (5 minutes): re-queues SQS with 15s delay and current executionCount - If max time exceeded: cancels all remaining PENDING/RUNNING jobs via cancelBigQueryJob() 5. For CREATESEGMENT / REBUILDSEGMENT actions: - Maps BigQuery status to app status (DONE/PENDING/RUNNING/CANCELED/ERROR) - For REBUILDSEGMENT: always sets queryStatus to DONE (not from BQ status mapping) - For CREATESEGMENT: sets queryStatus from BQ status mapping - Updates Segments table with totalBytesBilled (in GB, 0 for INSERT queries), buildAt - Clears Redis caches: redissegments{accountId}, {accountId}reportsegment{segmentId}, {accountId}reportsegments - Updates BigQueryJobs table with status and bigQueryJobId 6. Sends Pusher notification to frontend (channel-{accountId}) 7. Cascade rebuild: After successful completion (queryStatus = DONE), in parallel (Promise.all): - For all rebuild segments: clears segmentrebuildqueue:{segmentId} Redis key - For segments with isEmbeddedBy parents: calls triggerCascadeRebuild() which makes HTTP POST to {REACTAPP_API}/internal/segment/cascade-rebuild with { segmentId, cascadeId } to trigger next level via the main app's triggerCascadeRebuild - DB update, Pusher, and queue key cleanup are all handled in the lambda — no handleRebuildComplete round-trip to the main app

• Key files: models/segments.mjs (find, findByIds, update), models/bigqueryJobs.mjs, lib/bigquery.mjs, pusher.mjs, redis/index.mjs, helpers.mjs, constants/index.mjs
• Action types: create-segment, rebuild-segment, query-report, query-report-with-segment

3. ReportRequests (Supporting)

• Path: lambda-functions/ReportRequests/
• Purpose: Orchestrates BigQuery queries for reports with segment filtering
• Segment integration: Handles QUERYREPORTWITH_SEGMENT action, creates temporary tables from segment queries, returns segmentTables and comparisonSegmentTables

4. ExportReport (Supporting)

• Path: lambda-functions/ExportReport/
• Purpose: Exports report data to CSV/Excel with segment support
• Segment integration: Generates segment-specific column headers (S1:Name, S2:Name), creates per-segment metric columns

SQS Queues

Client-Side Architecture

Pages & Components

Segment Management Page:

• client/src/components/cms/subscriber/analytics/data-settings/manager-segment/index.js - Entry point (ManageSegment)
• client/src/components/cms/subscriber/analytics/data-settings/manager-segment/segment.js - Main listing with search, create button, default vs custom sections
• client/src/components/cms/subscriber/analytics/data-settings/manager-segment/ListSegment.js - SegmentItem rows with edit/delete/toggle, pagination (10/page)

Segment Creation/Editing:

• client/src/components/cms/subscriber/analytics/data-settings/manager-segment/SegmentDetail.js - Multi-step wizard with SegmentDetailContext

- Step 1 (SegmentRuleStep.js): Configure rule groups with AND/OR logic - Step 2 (SegmentInfoStep.js): Name and description (max 200 chars, unique)

Rule Condition Components (in steps/ directory):

Segment Usage in Reports:

• client/src/components/cms/subscriber/analytics/segments/SegmentDropdown.js - Segment selector dropdown (All/Default/Custom tabs, search, sticky toggle)
• client/src/components/cms/subscriber/analytics/segments/AddSegments.js - Multi-segment management with drag-and-drop reordering (S1, S2, S3...)
• client/src/components/cms/subscriber/analytics/segment-chart/ChartReportSegment.js - Chart rendering for segment analysis
• PopupPreviewInsideSegment.js - Preview modal showing matching field values

Redux State Management

Actions (client/src/actions/subscriber.js):

• fetchSegmentRequest(accountId) - Fetch all segments
• setListSegment(segment) - Update segments in store
• setListSegmentReport(segmentReport) - Set report segment configs
• setListSegmentReload / setListSegmentApplyReload - Trigger reloads
• setIsDoneSegment / setIsReadyRebuildSegment - Track operation status

Action Types (client/src/actions/types.js): SETLISTSEGMENT, SETLISTSEGMENTREPORT, SETLISTSEGMENTAPPLYRELOAD, SETLISTSEGMENTRELOAD, SETLISTSEGMENTDB, SETDBSEGMENTSELECTED, SETISLOADINGADDSEGMENTCHART, SETISLOADINGADDSEGMENTMETRIC, SETISDONESEGMENT, SETISREADYREBUILDSEGMENT, PUSHERUPDATE_SEGMENT

Reducer (client/src/reducers/subscriber.js): Manages listSegment, listSegmentReport, listSegmentDB, segment job statuses (in-progress, done, failed, canceled), totalBytesBilled, buildAt, queryStatus.

Client API Endpoints

Client Constants

File: client/src/constants/segment.js

Contains all SEGMENTOBJECT, CONDITIONFOCUS, SEGMENTFIELDTYPE, SEGMENTCONDITIONTYPE enums, condition operators, date range options, section categories for property grouping, and comprehensive PROPERTYMETRICSEGMENT_OPTIONS mappings for each object type.

End-to-End Flows

Segment Creation Flow

[User clicks "CREATE A NEW Segment"]
       |
       v
SegmentDetail modal opens (SegmentDetailContext)
       |
       v
Step 1: SegmentRuleStep
  - Select segment object (Session, User, People, Company, etc.)
  - Add rule groups with AND/OR conjunction
  - For each rule: select conditionFocus -> configure conditions
  - Preview field values via PopupPreviewInsideSegment
       |
       v
Step 2: SegmentInfoStep
  - Enter name (required, max 200 chars, unique)
  - Enter description (optional)
  - Click Save
       |
       v
POST /client/segment
       |
       v
Service: handleCreateSegment()
  - Generate UUID
  - Extract embedded segment IDs from ruleGroups
  - Validate embedded segments (circular deps, duplicates, self-embed)
  - Call getSegmentData() to build:
    - createTableQuery (CREATE OR REPLACE TABLE)
    - cronjobQuery (INSERT INTO for incremental updates)
  - Queue BigQuery job via QueueService.sendQueueCreateBigQueryJob()
  - Serialize ruleGroups, insert into Segments table
  - Sync isEmbeddedBy tracking on embedded segments
  - Clear Redis cache (REDIS_KEYS.SEGMENTS)
       |
       v
SQS: AWS_QUEUE_CHECK_BIG_QUERY_JOB
       |
       v
HandleCheckBigQueryJob Lambda
  - Poll BigQuery API every 15s (max 5 min)
  - On completion: update Segments table, clear Redis, send Pusher notification
       |
       v
[Frontend receives Pusher event, updates UI with queryStatus: 'done']

Segment Application to Reports Flow

[User opens report -> clicks Add Segment]
       |
       v
SegmentDropdown opens (All/Default/Custom tabs)
  - Search and select segments (max 4)
       |
       v
PUT /client/segment-report
       |
       v
Service: handleUpdateSegmentReport()
  - Create/update SegmentReport records
  - Enforce max 4 concurrent segments
  - Handle sticky behavior across reports
       |
       v
[Report query includes segment CTEs as filters]
  - getSegmentData() returns cteQuery + filterCondition
  - Report query JOINs segment table to filter results

Segment Cronjob (Incremental Update) Flow

[Scheduled SQS trigger → HANDLE_SEGMENT_BIG_QUERY queue]
       |
       v
HandleSegmentBigQuery Lambda (Initial Invocation)
  - Query PostgreSQL for active segments (isBlocked=false, queryStatus=done, <10GB)
  - getProcessingOrder() → topological sort into dependency levels
  - processNextLevel(level=0):
    - Filter out segments with failed dependencies
    - For each segment: replace date placeholders, submit BigQuery batch job
    - Last level: submit and return (no wait)
    - Other levels: self-queue SQS with 180s delay
       |
       v (SQS self-trigger after 180s)
HandleSegmentBigQuery Lambda (Continuation Invocation)
  - checkPendingJobs() → check BigQuery job statuses
  - If still running & checkCount < 10: re-queue with checkCount++
  - If all done or checkCount >= 10: processNextLevel(level+1)
  - Repeat until all levels processed
       |
       v
processSSTriggers()
  - Read S3 trigger rules, check schedule
  - If targets exist & hour < 12: send SQS to schedule report queue

Redis Caching

Segment Lifecycle

1. Creation: User defines rules -> validate embedded segments -> BigQuery table created -> sync isEmbeddedBy -> queryStatus: in-progress -> done
2. Caching: All segments cached in Redis per account (1-hour TTL)
3. Updating: Rules modified -> validate embedded segments -> BigQuery table recreated -> sync isEmbeddedBy (only if changed) -> cache cleared
4. Application: Applied to reports (max 4) -> used as CTE filters in report queries
5. Incremental Update: Cronjob queries INSERT new matching data into existing tables (processed in topological order for embedded dependencies)
6. Preview: User previews conditions -> BigQuery queried for sample/popular/unique values
7. Expiration/Rebuild: Full refresh via expire or rebuild -> drops and recreates table -> cascade rebuild triggers parent segments
8. Deletion: Block if embedded by others -> clean up isEmbeddedBy on child segments -> DB record deleted -> BigQuery table dropped -> cache cleared

Embed Segment Inside Another

Segments can be composed by embedding one segment into another segment's rule group. This narrows the rule group's results to only records that also belong to the embedded segment.

How It Works

Each rule group can optionally have an embeddedSegmentId field referencing another segment. The query builder adds a subquery filter:

• With regular conditions: WHERE ... AND objectId IN (SELECT objectId FROM segment_{embeddedId})
• Without conditions (embed only): SELECT DISTINCT objectId FROM segment_{embeddedId})

Bidirectional Tracking

Safety Mechanisms

Cascade Rebuild Flow

Segment B (embedded) is updated/rebuilt
       |
       v
triggerCascadeRebuild(segmentB.id)  [main app]
  - Check Redis visited set for cycle protection
  - Read segmentB.isEmbeddedBy -> [segmentA.id, segmentC.id]
       |
       v
For each parent (in parallel via Promise.all):
  queueSegmentRebuild(parentId)  [main app]
    - Set queryStatus to IN_PROGRESS + Pusher notification
    - Build new BigQuery table
    - Queue HandleCheckBigQueryJob
       |
       v
HandleCheckBigQueryJob Lambda completes:
  - Set queryStatus to DONE in DB
  - Send Pusher notification
  - Clear segment_rebuild_queue:{segmentId}
  - If segment has parents:
    - HTTP POST to main app /internal/segment/cascade-rebuild
    - triggerCascadeRebuild(parentId) -> rebuilds grandparents

Client-Side Flow

SegmentRuleStep
  |-- State: availableSegmentsForEmbed, isLoadingEmbedSegments, alreadyEmbeddedIds
  |-- Fetches: GET /client/segments/:accountId/available-for-embedding/:segmentId
  |-- Syncs queryStatus from Redux listSegment to available segments
  |-- Validates: no self-embed, no duplicates, selection required if toggled on
  |
  +-- CEmbeddedSegment (per rule group)
       |-- Checkbox: "Embed another segment into this rule group"
       |-- Dropdown: Select segment (filters out already-embedded, failed, in-progress)
       |-- Warning: Shows if selected segment is still building

Debugging Guide

Segment not building (stuck in-progress)

1. Check BigQuery job status: Look at HandleCheckBigQueryJob logs for the segment's job
2. Check SQS delivery: Verify AWSQUEUECHECKBIGQUERY_JOB received the message
3. Check Pusher: Verify notification was sent to channel-{accountId}
4. Check queryStatus in DB: Query Segments table for the segment's queryStatus, totalBytesBilled
5. Files: HandleCheckBigQueryJob/index.mjs, server/services/segments.js

Segment query failed

1. Check query syntax: Use POST /client/segment-testing/ to validate the segment query
2. Check BigQuery errors: Look at HandleCheckBigQueryJob logs for ERROR status
3. Check rule groups: Deserialize ruleGroups from DB and verify conditions are valid
4. Files: server/services/segmentTesting.js, server/models/query-builders/common/segment/

Segment data not appearing in reports

1. Check SegmentReport table: Verify apply = true for the segment/report combination
2. Check max segments: Only 4 segments can be applied per report simultaneously
3. Check segment status: Verify status = true (enabled) and queryStatus = 'done'
4. Check Redis cache: Stale cache may need clearing (redissegments{accountId})

Segment bytes billed too high (>10 GB)

1. Check totalBytesBilled: Segments table, totalBytesBilled column
2. Threshold: SEGMENTBYTESBILLED_THRESHOLD = 10 GB
3. Impact: Segments exceeding threshold are excluded from cronjob processing
4. Resolution: Rebuild segment or modify rules to reduce data scanned
5. Files: HandleSegmentBigQuery/constants/index.js, server/services/segments.js

Segment options not loading

1. Check Redis cache: Options are cached for 1 hour per variable type
2. Check variable type: Service fetches different data per SEGMENTVARIABLEOPTION type
3. Files: server/services/segments.js -> handleGetOptionDataByAccount()

Segment preview showing no data

1. Check BigQuery tables: Ensure the account's analytics tables have data
2. Check Redis cache: Preview results cached for 1 hour
3. Files: server/services/segments.js -> handlePreviewInsideSegment()

Embedded segment issues

1. Cannot delete segment: Check isEmbeddedBy column — segment is embedded by other segments. Remove the embedding from parent segments first.
2. Circular dependency error on save: checkCircularDependency() BFS detected a cycle. Check the chain: A embeds B embeds C embeds A.
3. Cascade rebuild not triggering: Check segmentSync.js logs (filter by [CascadeRebuild] prefix). Verify isEmbeddedBy array is populated. Check Redis queue key (segmentrebuildqueue:{segmentId}) and visited set (segmentcascadevisited:{cascadeId}).
4. Segments building in wrong order (cronjob): Check dependencyGraph.js topological sort. Verify getProcessingOrder() returns correct levels. Look for cycle fallback in logs.
5. Available segments list empty: handleGetAvailableSegmentsForEmbedding() filters out non-custom, self, and circular-dependency-causing segments. Verify account has other custom segments with queryStatus = 'done'.
6. Files: server/services/segments.js, server/services/segmentSync.js, HandleSegmentBigQuery/utils/dependencyGraph.js

Real-time update not reaching frontend

1. Check Pusher: HandleCheckBigQueryJob sends to channel-{accountId} with event PUSHERUPDATESEGMENT
2. Check Redux reducer: PUSHERUPDATESEGMENT action should update segment in store
3. Files: HandleCheckBigQueryJob/pusher.mjs, client/src/reducers/subscriber.js

Key Constants File

File: server/constants/segment.js

Contains all enums: SEGMENTOBJECT, CONDITIONFOCUS, SEGMENTFIELDTYPE, DEFAULTSEGMENTRULE, SEGMENTVARIABLEOPTION, and mapping objects (MAPPINGSEGMENTOBJECTTYPE, MAPPINGSEGMENTOBJECTID, MAPPINGCRMOBJECTTABLE, MAPPINGSFOBJECTTABLE, MAPPINGSALESCONVERSIONOBJECTTYPE, MAPPINGSALESCONVERSIONOBJECTID).

Key Environment Variables

• REDIS_URL - Redis connection for caching
• POSTGRESHOST, POSTGRESUSERNAME, POSTGRES_PASSWORD - PostgreSQL connection
• HANDLESEGMENTBIG_QUERY - SQS queue for segment cronjob processing (also used for self-queuing)
• AWSQUEUECHECKBIGQUERY_JOB - SQS queue for BigQuery job monitoring
• SENDMAILSCHEDULESAVEREPORTQUEUEURL - SQS queue for scheduled reports
• GOOGLEAPPLICATIONCREDENTIALS, GOOGLECLOUDPROJECT - BigQuery authentication
• PUSHERAPPID, PUSHERKEY, PUSHERSECRET, PUSHER_CLUSTER - Pusher real-time notifications
• S3BUCKETSTATIC - S3 bucket for server-side trigger configs