Systematic Memory Leak Hunter

You are a performance debugging specialist. Help me systematically find and fix a memory leak in my application:

**Symptoms:** [DESCRIBE — e.g., "Memory grows from 200MB to 2GB over 24 hours, then OOM kills the process"]
**Runtime:** [RUNTIME — e.g., "Node.js 20 / Python 3.12 / Java 21 / Go 1.22"]
**Application type:** [TYPE — e.g., "Express API server / React SSR / Background job processor"]
**Already tried:** [WHAT_YOU_TRIED — e.g., "Restarted the process, checked for obvious event listener leaks"]

Guide me through this debugging process:

**Phase 1: Reproduce**
- Exact steps to create a reproducible leak scenario
- How to measure baseline memory (commands and tools)
- Load generation script to accelerate the leak

**Phase 2: Capture**
- How to take heap snapshots at intervals (exact commands)
- How to enable allocation tracking without significant overhead
- What metrics to monitor during the capture window

**Phase 3: Analyze**
- How to compare heap snapshots to find growing objects
- How to identify the retainer chain (what's keeping objects alive)
- Common leak patterns specific to my runtime:
  - [Runtime-specific patterns, e.g., closures over large scopes, event emitter accumulation, timer references, global caches without eviction]

**Phase 4: Fix**
- Code pattern for each identified leak type with before/after examples
- How to verify the fix (expected memory profile shape)
- Preventive patterns to avoid similar leaks in the future

**Phase 5: Monitor**
- Metrics to set up for ongoing leak detection
- Alert thresholds and escalation rules
- Periodic heap snapshot automation for early detection

For each phase, provide exact commands, code snippets, and expected output.