PipeCD Introduction -English-

Slide 1

Slide 1 text

Continuous Delivery for Declarative Kubernetes, Serverless and Infrastructure Applications Developer Productivity Div, CyberAgent, Inc. Presents

Slide 2

Slide 2 text

Importance of CD 2 Commit Changes Build Test Staging Production Continuous Integration(CI) Continuous Delivery(CD) CD Automated Release DX Improvement Rapid and frequent releases Some elements that can be achieved with a good delivery process

Slide 3

Slide 3 text

CD Maturity 3 ● Deployment is not well procedural ● Deployment is done manually Crawl ● Deployment is procedural ● Easy to deploy with scripts and UI Walk ● Deployment data is versioned ● Sensitive data is kept secret ● The pipeline is sophisticated, and all delivery process is fully automated Run ● More secure deployment ● Rolling back quickly ● Monitoring and analyzing deployment Fly As CD proﬁciency increases, developers can focus more on the essence and reduce the risk of release.

Slide 4

Slide 4 text

The eﬀect of CD maturity 4 Conventional Coding Build/Delivery QA/Test Production/Verify Sophisticated Not only delivery time but also eﬃciency of the entire release process can be expected. Short lead times and number of releases make software development competitive

Slide 5

Slide 5 text

We need a more sophisticated delivery process ● Flexibility to handle complex architectures ○ Container Orchestration, Service Mesh ○ Serverless Architecture ● Release Safety ○ Approve ○ Canary Release, Blue-Green Deployment, etc... ● Secure ○ A completed CD in a user managed network ○ Conﬁdentiality ● Low Cost ○ Low learning costs by simple concepts and UI ○ Inexpensive running cost 5

Slide 6

Slide 6 text

Slide 7

Slide 7 text

Simple UI and visibility 7 The UI visualizes the state of the application in real time and clearly shows what happened at what timing

Slide 8

Slide 8 text

Great Automation 8 Rolling back based on error rate Conﬁguration Drift Detection

Slide 9

Slide 9 text

Self-hosted Runner Model 9 Deployment is operated by piped in the cluster, and application credentials are not leaked to the outside Excellent maintainability and low cost operation

Slide 10

Slide 10 text

Multi-provider & Multi-tenancy 10 Supports various platform, applications, and telemetry. Can be operated with multi-cluster tenancy.

Slide 11

Slide 11 text

Integration with Kubernetes ecosystems 11

Slide 12

Slide 12 text

Planning deployment types in the future 12 Cloud Run AWS Lambda [Alpha] [Alpha] [Incubating] Beta: Usable in production. Documented. Alpha: Demo-able, works end-to-end but has limitations. No guarantees on backward compatibility. Incubating: Under planning/developing the prototype and still not ready to be used. Amazon Elastic Container Service(ECS) [Incubating] [Incubating]

Slide 13

Slide 13 text

Core features of Control Plane ● [Beta] Sync and cancel deployment with Web UI ● SSO ○ [Beta] GitHub & GitHub Enterprise SSO ○ [Incubating] Google, BitBucket ● Self-Hosted Control Plane ○ [Beta] GCP Stack(GKE + Firestore + GCS) ○ [Incubating] AWS Stack(EKS + DynamoDB + S3) ○ [Alpha] Custom(Self-Managed k8s Cluster + MongoDB+ MinIO etc) ● Notiﬁcations ○ [Beta] Slack ○ [Incubating] Webhook ● [Incubating] ImageWatcher 13 Beta: Usable in production. Documented. Alpha: Demo-able, works end-to-end but has limitations. No guarantees on backward compatibility. Incubating: Under planning/developing the prototype and still not ready to be used.