FluxCD for continuous delivery in Kubernetes - Step-by-Step Execution

Process Flow - FluxCD for continuous delivery

Start: Developer pushes code to Git

↓

FluxCD detects Git changes

↓

FluxCD pulls updated manifests

↓

FluxCD applies manifests to Kubernetes cluster

↓

Cluster state updated to match Git

↓

Monitor for drift and repeat

FluxCD watches your Git repository for changes, then updates your Kubernetes cluster to match those changes automatically.

Execution Sample

Kubernetes

flux reconcile source git my-app
flux reconcile kustomization my-app
kubectl get pods

These commands tell FluxCD to sync the Git source and apply the Kubernetes manifests, then check the running pods.

Process Table

Step	Action	Input/Condition	Result/Output
1	Developer pushes code	New manifests committed to Git	Git repository updated
2	FluxCD detects change	Git repo has new commit	Triggers reconciliation
3	FluxCD pulls manifests	Fetch latest manifests from Git	Local cache updated
4	FluxCD applies manifests	Manifests describe deployment	Kubernetes cluster updated
5	Check cluster state	kubectl get pods	Pods reflect new deployment
6	Monitor for drift	Cluster state vs Git state	If drift, repeat sync

💡 FluxCD stops after cluster matches Git state and no drift detected

Status Tracker

Variable	Start	After Step 1	After Step 3	After Step 4	Final
Git Repository	Old manifests	New manifests committed	New manifests fetched	N/A	New manifests
FluxCD Cache	Empty or old	N/A	Updated with new manifests	N/A	Updated
Kubernetes Cluster	Old deployment	Old deployment	Old deployment	Updated deployment applied	Updated deployment
Pods	Old pods running	Old pods running	Old pods running	New pods starting	New pods running

Key Moments - 3 Insights

Why does FluxCD pull manifests from Git instead of pushing changes directly?

FluxCD follows GitOps principles where Git is the single source of truth. It pulls changes to ensure the cluster matches Git exactly, as shown in execution_table step 3.

What happens if the cluster state drifts from Git manifests?

FluxCD detects drift and automatically re-applies manifests to fix it, as described in execution_table step 6.

Why do we check pods after applying manifests?

Checking pods confirms that the new deployment is running correctly, verifying the cluster update in execution_table step 5.

Visual Quiz - 3 Questions

Look at the execution table, at which step does FluxCD update its local cache with new manifests?

A Step 2

B Step 3

C Step 4

D Step 5

💡 Hint

Refer to the 'Result/Output' column in step 3 describing 'Local cache updated'

According to the variable tracker, what is the state of the Kubernetes cluster after step 4?

A Old deployment

B New pods running

C Updated deployment applied

D Empty cluster

💡 Hint

Check the 'Kubernetes Cluster' row under 'After Step 4' in variable_tracker

If the developer does not push new code, what will FluxCD do according to the concept flow?

A Detect no changes and do nothing

B Apply old manifests repeatedly

C Delete all resources in cluster

D Crash due to no input

💡 Hint

Look at the concept_flow where FluxCD acts only when it detects Git changes

Concept Snapshot

FluxCD automates continuous delivery by syncing Kubernetes cluster state with Git repository.
It watches Git for changes, pulls updated manifests, and applies them to the cluster.
Git is the single source of truth; cluster state matches Git exactly.
FluxCD monitors for drift and re-applies manifests if needed.
Commands: 'flux reconcile source git' and 'flux reconcile kustomization' trigger sync.
Check cluster state with 'kubectl get pods' after deployment.

Full Transcript

FluxCD is a tool that helps keep your Kubernetes cluster in sync with your Git repository. When a developer pushes new code or configuration files to Git, FluxCD notices the change. It then pulls the updated files and applies them to the Kubernetes cluster. This process ensures the cluster always matches what is defined in Git. After applying changes, you can check the cluster's pods to see the new deployment running. FluxCD also watches for any differences between the cluster and Git, fixing them automatically. This approach is called GitOps, where Git is the single source of truth for your deployments.