Canary deployment pattern in Docker - Time & Space Complexity

Time Complexity: Canary deployment pattern

O(n)

Understanding Time Complexity

We want to understand how the time to deploy changes grows when using the canary deployment pattern with Docker.

Specifically, how does adding more users or instances affect deployment time?

Scenario Under Consideration

Analyze the time complexity of the following Docker commands for a canary deployment.

# Deploy new version to 1 container (canary)
docker run -d --name app_1 myapp:v2

# Wait and monitor canary container
# If stable, deploy to rest of containers
for i in $(seq 2 10); do
  docker stop app_$i
  docker rm app_$i
  docker run -d --name app_$i myapp:v2
  sleep 10
  # monitor logs or health
 done

This script deploys a new version first to one container, then sequentially updates the rest after monitoring.

Identify Repeating Operations

Look for repeated steps in the deployment process.

• Primary operation: Stopping, removing, and starting containers one by one.
• How many times: The loop runs for each container after the canary, here 9 times.

How Execution Grows With Input

As the number of containers increases, the deployment time grows roughly in direct proportion.

Input Size (n)	Approx. Operations
10	~10 container updates
100	~100 container updates
1000	~1000 container updates

Pattern observation: Deployment time grows linearly as more containers are updated one after another.

Final Time Complexity

Time Complexity: O(n)

This means deployment time increases directly with the number of containers to update.

Common Mistake

[X] Wrong: "Deploying to multiple containers at once will always be faster and have constant time."

[OK] Correct: Even if started simultaneously, monitoring and ensuring stability often requires sequential checks, so total time usually grows with the number of containers.

Interview Connect

Understanding how deployment time scales helps you design better release strategies and explain trade-offs clearly in real projects.

Self-Check

"What if we deployed all containers in parallel without waiting for monitoring? How would the time complexity change?"