metrics-collection

Parallel Metrics Collection from App, Infra, and Business Sources

You need a single dashboard showing request rate alongside CPU usage and revenue numbers, but those metrics live in three different systems. Collecting them sequentially triples the pipeline time. This workflow uses FORK_JOIN to collect application, infrastructure, and business metrics in parallel, then aggregates all 95 metrics from 3 sources.

Workflow

environment, timeRange
         |
         v
    FORK_JOIN ---------------------------+
    |                |                   |
    v                v                   v
+------------------+ +------------------+ +----------------------+
| mc_collect_app   | | mc_collect_infra | | mc_collect_business  |
+------------------+ +------------------+ +----------------------+
  45 metrics          32 metrics           18 metrics
  requestRate=1200    cpuUsage=65%         revenue=$24500
  errorRate=0.5%      memoryUsage=72%      orders=340
  p99Latency=230ms    diskIO=340           conversionRate=3.2%
    |                |                   |
    +------ JOIN ---+-------------------+
                |
                v
      +----------------+
      | mc_aggregate   |   totalMetrics: 95, sources: 3
      +----------------+

Workers

CollectAppMetrics -- Collects 45 application metrics from the specified environment. Key values: requestRate: 1200, errorRate: 0.5, p99Latency: 230.

CollectInfraMetrics -- Collects 32 infrastructure metrics. Key values: cpuUsage: 65, memoryUsage: 72, diskIO: 340.

CollectBusinessMetrics -- Collects 18 business metrics. Key values: revenue: 24500, orders: 340, conversionRate: 3.2.

AggregateMetrics -- Sums counts from all 3 sources: totalMetrics: 95, sources: 3, aggregatedAt: "2026-03-08T06:00:00Z".

Tests

29 unit tests cover each collector and the aggregation step.

Running

See ../../RUNNING.md for setup and execution instructions.