## TL;DR

PUT a combined service that does not exist and it is created; PUT one that exists and it is updated. That makes PUT the idempotent choice for infrastructure-as-code loops: you can re-apply the same desired state without a create-then-update dance.

## Steps

1. [official reference, real gotchas]: three Northflank behaviors that read backwards the first time. 1. put.* methods UPSERT. PUT a combined service that does not exist and it is created; PUT one that exists and it is updated. That makes PUT the idempotent choice for infrastructure-as-code loops: you can re-apply the same desired state without a create-then-update dance. 2. Pause is not delete. A paused service stops compute billing but keeps its config, secrets, and volumes. If your cleanup automation pauses instead of deleting, you are still paying for storage and the resource still exists. 3. Status COMPLETED means the deployment finished rolling out and the container is up and serving. It does NOT mean the process exited. When polling a service after deploy, COMPLETED is the success state you wait for, not a signal to tear down.

## When to use

You are seeing this: [official reference, real gotchas]: three Northflank behaviors that read backwards the first time. Use this skill when you run into "Northflank: put upserts, pause is not delete, COMPLETED means serving".

## When not to use

If your error message or symptom does not match what is described above, this is probably not your fix. Search for your exact error text instead of forcing this one to fit.

## Versions

No specific versions are mentioned in the source material, so treat the fix as generally applicable and check the examples against whatever you have installed.

## Why this happens

The original report does not dig into a root cause. It documents the symptom and the fix that resolved it.