Skip to content

PRODUCTION.md

Latest commit

 

History

History
51 lines (36 loc) · 1.74 KB

PRODUCTION.md

File metadata and controls

51 lines (36 loc) · 1.74 KB
Raw

Fork-Join — Production Deployment Guide

Required Environment Variables

Variable Required Description
CONDUCTOR_SERVER_URL Yes Conductor server API endpoint

No API keys required. This example uses in-memory data.

Architecture

The fork-join pattern executes three independent branches in parallel:

          +-- GetProductWorker ----+
          |                        |
Input --> +-- GetInventoryWorker --+--> MergeResultsWorker --> Output
          |                        |
          +-- GetReviewsWorker ----+

Each branch operates on its own data source with NO shared mutable state.

Error Handling

  • All workers validate productId input — missing/blank productId is terminal (FAILED_WITH_TERMINAL_ERROR)
  • GetProductWorker fails terminally for unknown product IDs — no synthetic fallback data
  • MergeResultsWorker requires ALL three branch outputs — if any branch fails, the merge fails
  • In Conductor, a FORK_JOIN with failOnError=true will terminate remaining branches if one fails

Security

  • Input validation prevents injection through product IDs
  • All data is in-memory; no external service calls
  • For production with real APIs: use circuit breakers and rate limiting on each branch

Deployment

export JAVA_HOME=/opt/homebrew/opt/openjdk@21/libexec/openjdk.jdk/Contents/Home
mvn clean package -DskipTests
java -jar target/fork-join-1.0.0.jar

Monitoring

  • Monitor individual branch task durations to identify slow branches
  • The merge step latency equals the slowest branch (parallel execution)
  • Alert on fj_get_product failures (indicates catalog data issues)
  • Alert on fj_merge_results failures (indicates broken branch contracts)