dotMemory vs. Alternatives: Why JetBrains Stands Out for .NET Profiling

10 dotMemory Tips Every .NET Developer Should Know

1. Start with the right profiling mode
   Use Sampling for low overhead and quick insights, Tracing when you need exact allocation call stacks, and Memory traffic to analyze allocations across threads and time.

2. Profile in an environment close to production
   Reproduce realistic workloads, input sizes, and runtimes to surface real memory issues that synthetic or tiny test runs may hide.

3. Collect multiple snapshots
   Take snapshots at key moments (startup, after heavy activity, before shutdown) and compare them to find growth patterns and leaks.

4. Use automatic inspections
   Enable dotMemory’s inspections to get instant hints (e.g., large objects, excessive Gen2 retention, finalizer queue growth) and prioritized suspects.

5. Filter and group objects smartly
   Group by type, namespace, assembly, or allocation stack to reduce noise and zero in on problematic object families.

6. Analyze retained sizes, not just instance counts
   A few large retained objects can matter more than many small short-lived instances—use retained size to find objects keeping others alive.

7. Inspect roots and object paths
   Use “Paths to GC roots” to see why objects aren’t collected; follow reference chains to the exact root preventing collection (static fields, event handlers, caches).

8. Use the Dominators view
   Dominators show which objects are responsible for most retained memory—fixing a dominator often yields the biggest memory win.

9. Watch for finalizer and native memory issues
   Monitor the finalizer queue and native allocations; unmanaged leaks or undisposed IDisposable implementations won’t show up as managed object growth alone.

10. Automate baseline comparisons in CI
    Integrate dotMemory Unit or CI checks to catch regressions early—set thresholds for heap size, allocation rate, or retained memory to fail builds when limits are exceeded.

Bonus quick commands:

• Enable “Show allocation call stacks” when tracing to link allocations to code.
• Use snapshots’ export to share findings with teammates.