pm2 v7.0.4 release notes

pm2 v7.0.4

v7.0.4Aug 24, 2026 · 8 days ago

Added 1

Start .ts apps with Node.js native type stripping when bun is not installed (Node.js >= 22.18 / 23.6, --experimental-strip-types auto-injected on 22.6+) with ts-node fallback now resolved from the app's own dependencies

Changed 1

Pin OpenTelemetry package versions in pm2 install-otel and re-emit legacy HTTP span tags (http.method, http.status_code, http.target) dropped by @opentelemetry/instrumentation-http >= 0.220

Fixed 7

Fix pm2 start --container / --container --dist crashing with Cannot find module due to wrong require depth in Containerizer.js
Bump js-yaml 4.3.0 → 4.3.1
Fix overlapping reloads colliding on the _old_<pm_id> slot and orphaning a cluster worker by refusing a reload while one is in progress
Ignore stale exit events from a replaced process to fix double start on Windows with shutdown_with_message
Fix retrying in NaNms kill log and ~1ms poll spam when kill_retry_time is unset by falling back to KILL_RETRY_TIME constant (100ms, overridable via PM2_KILL_RETRY_TIME)
Fix pm2 start/restart RPC hanging forever when a cluster worker dies before its online event by having executeApp conclude on exit-before-online
Surface the daemon's actual error in CLI output instead of masking everything as Process not found

Removed 1

Remove dead code including promise.min.js polyfill, IsAbsolute.js, unused Java/Ruby Dockerfile templates, and always-false win64 platform checks

From pm2

Start .ts apps with Node.js native type stripping when bun is not installed (Node.js >= 22.18 / 23.6, --experimental-strip-types auto-injected on 22.6+); ts-node fallback now resolved from the app's own dependencies

Fix pm2 start --container / --container --dist crashing with Cannot find module — wrong require depth in Containerizer.js from the v7 promptly internalization
Bump js-yaml 4.3.0 → 4.3.1
Fix overlapping reloads colliding on the _old_<pm_id> slot and orphaning a cluster worker — a reload is now refused while one is in progress #6129
Ignore stale exit events from a replaced process — fixes double start on Windows with shutdown_with_message #6142
Fix retrying in NaNms kill log and ~1ms poll spam when kill_retry_time is unset — fallback to KILL_RETRY_TIME constant (100ms, PM2_KILL_RETRY_TIME overridable)
Fix pm2 start/restart RPC hanging forever when a cluster worker dies before its online event (bad node_args, boot OOM) — executeApp now concludes on exit-before-online
Surface the daemon's actual error in CLI output instead of masking everything as Process not found
Pin OpenTelemetry package versions in pm2 install-otel and re-emit legacy HTTP span tags (http.method, http.status_code, http.target) dropped by @opentelemetry/instrumentation-http

= 0.220

Remove dead code: promise.min.js polyfill (native Promise), IsAbsolute.js (native path.isAbsolute), unused Java/Ruby Dockerfile templates, always-false win64 platform checks

pm2 v7.0.4