wrangler containers stop/start/restart/logs

[alexminza]

wrangler containers — missing instance lifecycle commands (restart, stop, start, logs, exec)

Problem

wrangler containers currently supports listing and deleting containers/instances, but has no commands for day-to-day instance lifecycle management:

• No restart — cannot restart a running container instance without deleting and redeploying
• No stop/start — cannot stop an instance without deleting it
• No logs — cannot stream or tail container instance logs (must use the dashboard)
• No exec/ssh — cannot execute commands inside a running instance for debugging

This forces operators to either use the Cloudflare dashboard UI or make raw API calls to Cloudchamber for routine operations like restarting a misbehaving instance or inspecting logs during an incident.

Proposed commands

Command	Description
wrangler containers restart <ID> [--instance <INSTANCE_ID>]	Restart a specific instance (or all instances) of a container application
wrangler containers stop <ID> [--instance <INSTANCE_ID>]	Stop a running instance without deleting the application
wrangler containers start <ID> [--instance <INSTANCE_ID>]	Start a previously stopped instance
wrangler containers logs <ID> [--instance <INSTANCE_ID>]	Stream/tail logs from a container instance (--follow, --tail N, --since)
wrangler containers exec <ID> --instance <INSTANCE_ID> -- <command>	Execute a command inside a running instance
wrangler containers redeploy <ID>	Force re-pull image and restart all instances (equivalent to restart with image refresh)

Current state (wrangler 4.76.0)

Existing wrangler containers subcommands:

wrangler containers list             # List containers
wrangler containers info <ID>        # Get container info
wrangler containers instances <ID>   # List instances
wrangler containers delete <ID>      # Delete a container
wrangler containers build <PATH>     # Build an image
wrangler containers push <TAG>       # Push image to CF Registry
wrangler containers registries       # Manage registries
wrangler containers images           # Manage images

The gap: you can create, list, inspect, and destroy — but you cannot operate on running instances.

Workaround

Currently the only way to "restart" a container instance is to delete and redeploy, which causes downtime and requires re-running the full deployment pipeline. For logs, operators must navigate to the Cloudflare dashboard manually.

Use case

Running long-lived services (e.g., AI gateway, bot backends) on Cloudflare Containers that occasionally need restarts due to upstream software bugs, configuration changes, or memory leaks — without triggering a full redeployment cycle.