Instance states (running, stopped, terminated) in GCP - Deep Dive

Overview - Instance states (running, stopped, terminated)

What is it?

Instance states describe the current condition of a virtual machine in the cloud. Common states include running, stopped, and terminated. Each state shows if the instance is active, paused, or deleted. Understanding these states helps manage cloud resources effectively.

Why it matters

Without knowing instance states, you might waste money by leaving machines running when not needed or lose data by terminating instances accidentally. Proper state management ensures cost control, resource availability, and system reliability in cloud environments.

Where it fits

Before learning instance states, you should understand what virtual machines are and basic cloud concepts. After this, you can learn about instance lifecycle management, automation, and cost optimization in cloud platforms.

Mental Model

Core Idea

An instance's state is like a traffic light showing whether the machine is active, paused, or gone.

Think of it like...

Think of a virtual machine like a car: running means the car is driving, stopped means the car is parked but ready to go, and terminated means the car is sold and no longer available.

┌─────────────┐     start     ┌─────────────┐
│  STOPPED    │─────────────▶│  RUNNING    │
└─────────────┘              └─────────────┘
       ▲                          │
       │          stop            │ terminate
       │                          ▼
┌─────────────┐              ┌─────────────┐
│ TERMINATED  │◀─────────────│  RUNNING    │
└─────────────┘              └─────────────┘

Build-Up - 7 Steps

1

Foundation

What is an Instance State

Concept: Introduce the idea that virtual machines have different states representing their activity.

A virtual machine (VM) in the cloud can be in different states. The main states are running, stopped, and terminated. Running means the VM is active and working. Stopped means the VM is powered off but still exists. Terminated means the VM is deleted and no longer available.

Result

You understand that instance states tell you if a VM is active, paused, or gone.

Knowing that instances have states helps you control when they use resources and cost.

2

Foundation

Running State Explained

Concept: Explain what happens when an instance is running.

When an instance is running, it is powered on and ready to perform tasks. It uses CPU, memory, and network resources. You can connect to it and run applications. Running instances incur charges because they consume resources.

Result

You can identify when a VM is active and understand it costs money while running.

Understanding running state helps you know when your cloud resources are actively working and billing starts.

3

Intermediate

Stopped State Details

🤔Before reading on: Do you think a stopped instance still uses cloud resources and costs money? Commit to your answer.

Concept: Describe what stopped means and its impact on resources and billing.

A stopped instance is powered off but still exists in your cloud account. It does not use CPU or memory, so it does not incur compute charges. However, storage costs for the instance's disk remain because the data is saved. You cannot connect to a stopped instance until you start it again.

Result

You know that stopping an instance pauses compute charges but keeps storage costs.

Knowing stopped state helps you save money by pausing compute without losing data.

4

Intermediate

Terminated State and Consequences

🤔Before reading on: Does terminating an instance keep its data and allow restarting later? Commit to your answer.

Concept: Explain what termination means and its effect on the instance and data.

Terminating an instance means deleting it permanently. The VM and its attached disks are removed unless you specifically keep the disks. You cannot restart a terminated instance. This frees all resources and stops all charges. However, any data not saved externally is lost.

Result

You understand that termination deletes the instance and stops all charges but risks data loss.

Knowing termination effects prevents accidental data loss and helps manage resource cleanup.

5

Intermediate

State Transitions and Lifecycle

Concept: Show how instances move between states and what triggers changes.

Instances move between states by user actions or system events. Starting a stopped instance moves it to running. Stopping a running instance moves it to stopped. Terminating an instance deletes it from any state. Some cloud platforms allow automatic state changes based on schedules or policies.

Result

You can predict how instance states change and control their lifecycle.

Understanding state transitions helps automate and optimize cloud resource usage.

6

Advanced

Billing Impact of Instance States

🤔Before reading on: Do you think stopped instances always cost the same as running ones? Commit to your answer.

Concept: Detail how billing works for each state in cloud platforms like GCP.

Running instances incur charges for compute resources and storage. Stopped instances do not incur compute charges but still incur storage charges for disks. Terminated instances do not incur any charges unless persistent disks are retained. Understanding billing helps optimize costs by choosing the right state.

Result

You can manage cloud costs by controlling instance states effectively.

Knowing billing differences prevents unexpected charges and supports cost-saving strategies.

7

Expert

Unexpected Behaviors in State Management

🤔Before reading on: Can a stopped instance lose its external IP address automatically? Commit to your answer.

Concept: Reveal subtle behaviors like IP address release and disk persistence during state changes.

In some cloud platforms, stopping an instance releases its ephemeral external IP address unless it is static. Also, some instance metadata or temporary storage may be lost on stop. Persistent disks remain intact. These behaviors affect connectivity and data availability after stopping and restarting.

Result

You anticipate and handle subtle state-related behaviors to avoid downtime or data loss.

Understanding these nuances helps design resilient and reliable cloud architectures.

Under the Hood

Instance states are managed by the cloud control plane, which tracks each VM's lifecycle. Running means the hypervisor allocates CPU, memory, and network resources to the VM. Stopped means the VM is powered off but its disk images remain stored. Terminated means the VM and its resources are deleted from the control plane and storage unless preserved separately.

Why designed this way?

This design balances resource efficiency and user control. Running instances consume resources and cost money. Stopped instances save compute costs but keep data for quick restart. Termination frees all resources to avoid waste. Alternatives like instant deletion on stop would risk data loss or slow restarts.

┌───────────────┐       ┌───────────────┐       ┌───────────────┐
│   Control     │──────▶│   Hypervisor  │──────▶│   Physical    │
│    Plane      │       │   Manages VM  │       │   Hardware    │
└───────────────┘       └───────────────┘       └───────────────┘
       │                       │                       │
       │                       │                       │
       ▼                       ▼                       ▼
┌─────────────┐         ┌─────────────┐         ┌─────────────┐
│  Instance   │         │  Disk Image │         │  Network    │
│   State     │         │  Storage    │         │  Resources  │
└─────────────┘         └─────────────┘         └─────────────┘

Myth Busters - 4 Common Misconceptions

Quick: Does stopping an instance always stop all charges? Commit to yes or no.

Common Belief: Stopping an instance stops all charges immediately.

Reality: Stopping an instance stops compute charges but storage charges for disks continue.

Why it matters: Ignoring storage costs can lead to unexpected bills even when instances are stopped.

Quick: Can you restart a terminated instance? Commit to yes or no.

Common Belief: Terminated instances can be restarted later like stopped ones.

Reality: Terminated instances are deleted and cannot be restarted; you must create a new instance.

Why it matters: Assuming restart is possible after termination can cause data loss and downtime.

Quick: Does stopping an instance always keep its external IP address? Commit to yes or no.

Common Belief: Stopping an instance always keeps the same external IP address.

Reality: Stopping an instance releases ephemeral external IPs unless they are static or reserved.

Why it matters: Losing IP addresses can break connections and require reconfiguration.

Quick: Are all instance states instantly reflected in the cloud console? Commit to yes or no.

Common Belief: Instance states update instantly and always reflect the true status.

Reality: There can be delays or temporary inconsistencies in state reporting due to system propagation.

Why it matters: Relying on immediate state changes can cause confusion or incorrect automation triggers.

Expert Zone

1

Stopping an instance may clear temporary local storage, so data should not be stored there if persistence is needed.

2

Some cloud providers allow 'suspend' or 'hibernate' states that save memory state to disk for faster restart, differing from stopped state.

3

Automated policies can trigger state changes based on usage patterns, enabling cost optimization without manual intervention.

When NOT to use

Using stopped state is not suitable when you need guaranteed IP address retention or uninterrupted service. In such cases, consider using reserved IPs or load balancers. Termination should be avoided if data persistence is required; instead, use snapshots or backups.

Production Patterns

In production, instances are often stopped during off-hours to save costs and started automatically when needed. Termination is used for scaling down or replacing faulty instances. Monitoring tools track instance states to trigger alerts or automation workflows.

Connections

State Machines

Instance states follow a state machine pattern with defined transitions.

Understanding state machines helps grasp how instance states change predictably and how to design automation around them.

Operating System Power States

Instance states are similar to OS power states like sleep, hibernate, and shutdown.

Knowing OS power states clarifies why stopped instances save data but do not consume CPU, like a computer in sleep mode.

Project Management Task Status

Instance states relate to task statuses like active, paused, or completed.

Seeing instance states as task statuses helps non-technical learners understand lifecycle and resource management concepts.

Common Pitfalls

#1 Assuming stopping an instance stops all charges.

Wrong approach:

Stop instance and ignore storage costs, expecting zero charges.

Correct approach:

Stop instance and monitor storage costs; delete or snapshot disks if not needed.

Root cause:

Misunderstanding that storage resources remain allocated and billed after stopping.

#2 Terminating an instance without backing up data.

Wrong approach:

Terminate instance directly after use without snapshots.

Correct approach:

Create snapshots or backups before terminating to preserve data.

Root cause:

Not realizing termination deletes instance and attached disks permanently.

#3 Relying on ephemeral IP addresses after stopping.

Wrong approach:

Stop instance and expect the same external IP to remain assigned.

Correct approach:

Reserve static IP addresses if persistent IPs are needed across stops.

Root cause:

Not knowing that ephemeral IPs are released on stop.

Key Takeaways

Instance states indicate whether a virtual machine is active, paused, or deleted, guiding resource use and cost.

Running instances consume compute resources and incur charges; stopped instances save compute costs but keep storage charges.

Terminated instances are deleted permanently and cannot be restarted, so data must be backed up beforehand.

State transitions allow control over instance lifecycle, enabling cost optimization and automation.

Understanding subtle behaviors like IP address release and storage billing prevents unexpected downtime and costs.