Zero‑downtime PM2 reload for stateful Node.js apps

Can PM2's reload command achieve zero-downtime migration for a small stateful Node.js application?

0 reputation · 31 Mar 2025, 13:21 UTC

110.8K views0

Goal: Migrate a small Node.js application to PM2 Cluster Mode to enable zero‑downtime reloads while maintaining service availability. Constraint: The application stores user session data in local memory, so a reload that replaces instances may lose that state unless handled. Uncertainty: It is unclear whether PM2’s reload mechanism can preserve in‑memory session data across sequential restarts without external storage or code changes. How can session state be preserved across PM2 reloads without introducing an external store?

2,706 following

Keep exploring this topic

Which shutdown strategy should a PM2 cluster-mode app use for WebSocket clients during zero-downtime reloads?

A Node.js service moving from PM2 fork mode to cluster mode needs a shutdown strategy for long-lived WebSocket clients before zero-downtime reloads are safe.

PM2 Master and Worker Processes: IPC Overhead During High-Frequency Metric Emission

Investigating if PM2's internal IPC channel, used for high-frequency logs and metrics, contributes to event-loop lag in cluster-mode workers during concurrent request processing.

PM2 cluster worker crash during shutdown handler: restart or abandon?

PM2's docs don't clarify whether a cluster worker that crashes during its shutdown handler gets restarted or is left dead before the master's kill_timeout expires. This affects how cleanup code must be written.

PM2 Cluster Mode and Database Schema Migration Interoperability

Exploring the risks of 'split-brain' scenarios when using PM2 zero-downtime reloads alongside destructive database schema migrations.

Can PM2's reload command achieve zero-downtime migration for a small stateful Node.js application?

0 answers

Your perspective belongs here.

0 question comments