Retention experiments
Force-GC cycles, survival ratios, and growth patterns that separate churn from sticky populations.
Ruby · MRI 3.1+ · Local-only · No SaaS
Evidence-driven diagnostics for object retention, heap growth, and leak-shaped patterns in long-running Ruby processes — without claiming a leak unless the evidence is strong.
Available on RubyGems ·
gem install heapscope
gem install heapscope
heapscope about
heapscope doctor
Snapshots, multi-cycle experiments, allocation sites, thread-locals, and CI budgets — all offline.
Force-GC cycles, survival ratios, and growth patterns that separate churn from sticky populations.
HS001–HS010 diagnostics separate facts, derived behavior, hypotheses, and suspected causes.
Optional middleware and matchers without pulling frameworks into the core gem.
No network, no telemetry, no object-value dumps by default. Reports stay on your machine.
CI presets (rails_request, sidekiq_job, ci_strict) plus ranked follow-ups on every report.
heapscope watch samples RSS and live slots and flags spike anomalies — still fully local.
Get the release on rubygems.org/gems/heapscope. Source and release notes live on GitHub Releases.
# Install
gem install heapscope
# Or Gemfile
gem "heapscope", "~> 0.6"
require "heapscope"
HeapScope.probe(title: "import") do
ImportJob.perform_now
end
Independent software — fund ongoing maintenance via GitHub Sponsors or thanks.dev.