Version Log
|
Date |
Author |
Version |
Change
Description |
|
04/13/2025 |
Srikanth Mamillapalli |
0.1 |
Initial Draft |
Table of Contents
|
# |
Strategy |
Downtime |
Rollback Speed |
Extra Infrastructure |
Complexity |
Common Usage |
|
1 |
Recreate Deployment |
High |
Medium |
No |
Low |
Small/Internal Apps |
|
2 |
Rolling Deployment |
None/Minimal |
Medium |
No |
Medium |
Kubernetes Default |
|
3 |
Blue-Green Deployment |
None |
Very Fast |
Yes |
Medium |
Banking, Payments |
|
4 |
Canary Deployment |
None |
Fast |
Partial |
High |
Large Scale Systems |
|
5 |
A/B Testing Deployment |
None |
Fast |
Partial |
High |
Product Experiments |
|
6 |
Shadow Deployment |
None |
Fast |
Yes |
High |
Performance Testing |
|
7 |
Feature Toggle (Feature Flag) |
None |
Instant |
No |
Medium |
Continuous Delivery |
|
8 |
Ring-Based Deployment |
None |
Fast |
No |
High |
Enterprise Applications |
|
9 |
Red-Black Deployment |
None |
Very Fast |
Yes |
Medium |
Similar to Blue-Green |
|
10 |
Progressive Delivery |
None |
Fast |
Partial |
High |
Cloud-Native Systems |
Quick Reference Table
1. Recreate Deployment
|
Category |
Details |
|
Strategy Name |
Recreate Deployment |
|
How It Works |
1. Stop Version 12. Deploy Version 23.
Start Version 2 |
|
Deployment Flow |
v1 Running → Application Down →
v2 Running |
|
Downtime |
Yes (Application is unavailable during
deployment) |
|
Rollback Speed |
Medium (Requires redeployment of previous
version) |
|
Infrastructure Requirement |
No additional infrastructure required |
|
Complexity |
Low |
|
Advantages |
• Very simple• Easy to implement• No
additional resources required |
|
Disadvantages |
• Downtime occurs during deployment• Risky
for production systems• Poor user experience during upgrades |
|
Kubernetes Configuration |
yaml strategy: type: Recreate |
|
Best For |
• Internal tools• Non-critical
applications• Development/Test environments |
|
Not Recommended For |
• Banking applications• Payment systems•
Customer-facing production applications |
|
Spring Boot Example |
Stop existing Spring Boot pod/container,
deploy new version, then start new pod/container |
|
Real-Time Use Case |
Internal Admin Portal, Reporting Dashboard,
QA/Test Environment |
|
Production Suitability |
Low |
|
Zero Downtime Support |
No |
Visual Representation
Before Deployment
-----------------
Spring Boot v1
|
Users
Deployment Starts
-----------------
Stop v1
|
Application Down
Deployment Completes
--------------------
Start v2
|
Users
Kubernetes YAML Example
apiVersion: apps/v1
kind: Deployment
metadata:
name:
user-service
spec:
strategy:
type:
Recreate
replicas: 1
template:
spec:
containers:
- name:
user-service
image: user-service:v2
2. Blue-Green Deployment
|
Category |
Details |
|
Strategy Name |
Blue-Green Deployment |
|
How It Works |
Maintain two identical environments and
switch traffic |
|
Deployment Flow |
Blue(v1) Active → Deploy Green(v2)
→ Test → Switch Traffic |
|
Downtime |
No |
|
Rollback Speed |
Very Fast |
|
Infrastructure Requirement |
Yes (Double Infrastructure) |
|
Complexity |
Medium |
|
Advantages |
• Instant rollback• Near-zero downtime•
Safer releases |
|
Disadvantages |
• Higher cost• Database compatibility
challenges |
|
Kubernetes Configuration |
Usually implemented using Services/Ingress
switching |
|
Best For |
Banking, Payments, Authentication Services |
|
Not Recommended For |
Budget-sensitive projects |
|
Production Suitability |
Very High |
|
Zero Downtime Support |
Yes |
Two environments exist simultaneously.
Traffic is switched to the new version after validation, enabling immediate
rollback.
3. Canary Deployment
|
Category |
Details |
|
Strategy Name |
Canary Deployment |
|
How It Works |
Release new version to a small subset of
users |
|
Deployment Flow |
10% → 25% → 50% → 75%
→ 100% |
|
Downtime |
No |
|
Rollback Speed |
Fast |
|
Infrastructure Requirement |
Partial |
|
Complexity |
High |
|
Advantages |
• Reduced risk• Real-user validation•
Gradual rollout |
|
Disadvantages |
• Requires advanced monitoring• Traffic
routing complexity |
|
Kubernetes Tools |
Istio, Argo Rollouts, OpenShift Routes |
|
Best For |
High-traffic systems |
|
Not Recommended For |
Small teams lacking monitoring tools |
|
Production Suitability |
Very High |
|
Zero Downtime Support |
Yes |
Canary deployments expose a small percentage of
users to the new version before wider rollout.
4. A/B Testing Deployment
|
Category |
Details |
|
Strategy Name |
A/B Testing |
|
How It Works |
Different users receive different versions |
|
Deployment Flow |
Group A → Version AGroup
B → Version B |
|
Downtime |
No |
|
Rollback Speed |
Fast |
|
Infrastructure Requirement |
Partial |
|
Complexity |
High |
|
Advantages |
• Business validation• User behavior comparison |
|
Disadvantages |
• Analytics required• Complex routing |
|
Best For |
UI/UX Experiments |
|
Not Recommended For |
Backend-only services |
|
Production Suitability |
High |
|
Zero Downtime Support |
Yes |
A/B testing helps compare user engagement and
business metrics across multiple versions.
5. Shadow Deployment
|
Category |
Details |
|
Strategy Name |
Shadow Deployment |
|
How It Works |
Mirror production traffic to new version |
|
Deployment Flow |
User → v1 ResponseTraffic
Copy → v2 |
|
Downtime |
No |
|
Rollback Speed |
Fast |
|
Infrastructure Requirement |
Yes |
|
Complexity |
High |
|
Advantages |
• Real traffic testing• No customer impact |
|
Disadvantages |
• Additional infrastructure cost• Increased
monitoring needs |
|
Tools |
Istio, Envoy |
|
Best For |
Performance validation |
|
Not Recommended For |
Cost-sensitive projects |
|
Production Suitability |
High |
|
Zero Downtime Support |
Yes |
Shadow deployment duplicates production traffic to
the new version without exposing users to it.
6. Feature Toggle (Feature
Flag)
|
Category |
Details |
|
Strategy Name |
Feature Toggle / Feature Flag |
|
How It Works |
Deploy code with feature disabled |
|
Deployment Flow |
Deploy → Feature OFF → Enable
Gradually |
|
Downtime |
No |
|
Rollback Speed |
Instant |
|
Infrastructure Requirement |
No |
|
Complexity |
Medium |
|
Advantages |
• Instant enable/disable• No redeployment
needed |
|
Disadvantages |
• Code complexity• Technical debt if not
cleaned up |
|
Tools |
LaunchDarkly, Unleash |
|
Best For |
Continuous delivery |
|
Not Recommended For |
Long-term permanent features |
|
Production Suitability |
Very High |
|
Zero Downtime Support |
Yes |
Features are deployed but hidden behind flags,
allowing controlled activation.
7. Ring-Based Deployment
|
Category |
Details |
|
Strategy Name |
Ring-Based Deployment |
|
How It Works |
Rollout in stages to predefined user groups |
|
Deployment Flow |
Ring0 → Ring1 → Ring2 →
Ring3 |
|
Downtime |
No |
|
Rollback Speed |
Fast |
|
Infrastructure Requirement |
No |
|
Complexity |
High |
|
Advantages |
• Controlled rollout• Early issue detection |
|
Disadvantages |
• Operational complexity |
|
Best For |
Enterprise SaaS Products |
|
Not Recommended For |
Small projects |
|
Production Suitability |
High |
|
Zero Downtime Support |
Yes |
|
Ring |
Users |
|
Ring 0 |
Developers |
|
Ring 1 |
Internal Employees |
|
Ring 2 |
Beta Customers |
|
Ring 3 |
All Customers |
8. Red-Black Deployment
|
Category |
Details |
|
Strategy Name |
Red-Black Deployment |
|
How It Works |
Same concept as Blue-Green using Red and
Black names |
|
Deployment Flow |
Red(v1) Active → Black(v2) →
Traffic Switch |
|
Downtime |
No |
|
Rollback Speed |
Very Fast |
|
Infrastructure Requirement |
Yes |
|
Complexity |
Medium |
|
Advantages |
• Immediate rollback• Zero downtime |
|
Disadvantages |
• Double infrastructure cost |
|
Best For |
Mission-critical applications |
|
Production Suitability |
Very High |
|
Zero Downtime Support |
Yes |
9. Progressive Delivery
|
Category |
Details |
|
Strategy Name |
Progressive Delivery |
|
How It Works |
Combination of Canary + Feature Flags +
Monitoring |
|
Deployment Flow |
1% → 5% → 10% → 50%
→ 100% |
|
Downtime |
No |
|
Rollback Speed |
Automatic/Fast |
|
Infrastructure Requirement |
Partial |
|
Complexity |
Very High |
|
Advantages |
• Safest deployment model• Automated
rollback |
|
Disadvantages |
• Requires mature DevOps ecosystem |
|
Tools |
Argo Rollouts, Flagger, Istio |
|
Best For |
Cloud-native microservices |
|
Production Suitability |
Very High |
|
Zero Downtime Support |
Yes |
Progressive delivery combines traffic control,
feature flags, monitoring, and automated rollback to safely release software.
10. Final Comparison Table
|
Strategy |
Downtime |
Rollback |
|
Cost |
Complexity |
Enterprise Usage |
|
Recreate |
High |
Medium |
|
Low |
Low |
Low |
|
Rolling |
None |
Medium |
|
Low |
Medium |
Very Common |
|
Blue-Green |
None |
Instant |
|
High |
Medium |
Banking/Payments |
|
Canary |
None |
Fast |
|
Medium |
High |
Large Scale Apps |
|
A/B Testing |
None |
Fast |
|
Medium |
High |
Product Teams |
|
Shadow |
None |
Fast |
|
High |
High |
Performance Testing |
|
Feature Flags |
None |
Instant |
|
Low |
Medium |
Continuous Delivery |
|
Ring-Based |
None |
Fast |
|
Medium |
High |
Enterprise SaaS |
|
Red-Black |
None |
Instant |
|
High |
Medium |
Mission Critical |
|
Progressive Delivery |
None |
Automatic |
|
Medium |
Very High |
Modern Cloud-Native |
1. Recreate Deployment
How It Works
Stop
Version 1
↓
Deploy Version 2
↓
Start Version 2
Example
v1
Running
↓
Application Down
↓
v2 Running
Advantages
Ø
Very
simple
Ø
Easy to
implement
Ø
No
additional resources
Disadvantages
Ø
Downtime
occurs
Ø
Risky for
production
Kubernetes
strategy:
type: Recreate
Best For
2. Rolling Deployment
How It Works
Replace
instances gradually.
Before
v1 v1 v1 v1
Step1
v2 v1 v1 v1
Step2
v2 v2 v1 v1
Step3
v2 v2 v2 v1
Final
v2 v2 v2 v2
Advantages
Ø
No
downtime
Ø
Resource
efficient
Disadvantages
Ø
Mixed
versions coexist
Ø
Rollback
takes time
Kubernetes
strategy:
type: RollingUpdate
Best For
Ø
Most
Spring Boot microservices
Ø
Standard
Kubernetes deployments
3. Blue-Green Deployment
How It Works
Production
|
Blue (v1)
Deploy v2
Green (v2)
Switch Traffic
Production
|
Green (v2)
Advantages
Ø
Near zero
downtime
Ø
Instant
rollback
Disadvantages
Ø
Double
infrastructure cost
Ø
Database
compatibility issues
Best For
Ø
Banking
applications
Ø
Payment
systems
Ø
Mission-critical
services
4. Canary Deployment
How It Works
Release
to a small percentage of users.
Users
|
Load Balancer
|
90% → v1
10% → v2
Gradually:
10%
25%
50%
75%
100%
Advantages
Ø
Lower risk
Ø
Real user
validation
Disadvantages
Ø Traffic management required
Ø Monitoring is critical
Best For
Ø
High
traffic systems
Ø
SaaS
products
5. A/B Testing Deployment
How It Works
User
Group A → Version A
User Group B → Version B
Example
India
Users → New UI
US Users → Old UI
Advantages
Ø
Business
validation
Ø
User behavior comparison
Disadvantages
Ø
Complex
routing
Ø
Additional
analytics required
Best For
Ø
Product
teams and UI experimentation
6. Shadow Deployment
How It Works
Mirror
production traffic.
Users
|
Version 1
|
Copy Traffic
|
Version 2
Users
only see Version 1 responses.
Advantages
Ø
Real
production traffic testing
Ø
No user impact
Disadvantages
Ø
Higher
infrastructure cost
Ø
More
monitoring needed
Best For
Ø
Performance
validation
Ø
Large
enterprises
7. Feature Toggle (Feature Flag)
How It Works
Deploy
Code
|
Feature Disabled
|
Enable for Users
Example
if(featureFlagEnabled){
showNewCheckout();
}
Advantages
Ø
Instant
enable/disable
Ø
No
redeployment
Disadvantages
Ø
Additional
code complexity
Ø
Technical
debt if flags are not removed
Best For Agile teams and
Continuous deployment
8. Ring-Based Deployment
Users
are grouped into rings.
Ring
0 → Internal Team
Ring 1 → Beta Users
Ring 2 → 10% Users
Ring 3 → All Users
Advantages
Ø
Safe
rollout
Ø
Early
issue detection
Disadvantages
Ø
Complex
management
Best For
Ø Enterprise SaaS platforms
9. Red-Black Deployment
Practically
same as Blue-Green.
Red
Environment → Live
Black Environment → New
Switch Traffic
Advantages
Ø
Instant
rollback
Ø
Zero
downtime
· Disadvantages
Ø
Double
infrastructure
· Best For
Ø
Large
production systems
10. Progressive Delivery
Combination
of:
Feature
Flags
+
Canary
+
A/B Testing
+
Monitoring
Flow
1%
Users
↓
10% Users
↓
50% Users
↓
100% Users
Automatic
rollback if errors increase.
Advantages
Ø Safest deployment model and Highly
automated
Ø Disadvantages Requires mature DevOps practices
Best For
Ø
Cloud-native
platforms
Ø
Microservices
ecosystems
Real-Time Enterprise Examples
|
Company Type |
Preferred Strategy |
|
Banking |
Blue-Green |
|
Payment Gateway |
Blue-Green + Canary |
|
E-Commerce |
Canary + Feature Flags |
|
SaaS Product |
Progressive Delivery |
|
Internal HR Portal |
Rolling Update |
|
Enterprise Microservices |
Rolling + Canary |
|
High Traffic Consumer Apps |
Canary |
|
Mission Critical Applications |
Blue-Green |
What deployment strategy do you use in
Kubernetes/OpenShift?
"For
regular Spring Boot microservices, we use Rolling Updates. For critical
services such as authentication, payments, or order processing, we use
Blue-Green deployments to achieve zero downtime and instant rollback. For
large-scale customer-facing applications, we combine Canary deployments with
Feature Flags and automated monitoring to reduce release risk."