Longhorn ArgoCD
You set up a longhorn argocd workflow by tracking storage manifests inside a Git repo. Then, the continuous delivery controller syncs those manifests to your Kubernetes cluster. Platform teams use this pattern to manage block storage across many nodes. Still, teams often grant the delivery controller broad cluster admin roles during setup. Broad roles let anyone with repo write rights change storage volumes or delete data. How a longhorn argocd workflow operates This workflow links your versi
Your coding agents run with your access.
Claude Code reaches prod with a developer's credentials, and approval prompts get clicked through. hoop.dev checks every query, masks sensitive data, and holds risky changes for a human.
You set up a longhorn argocd workflow by tracking storage manifests inside a Git repo. Then, the continuous delivery controller syncs those manifests to your Kubernetes cluster. Platform teams use this pattern to manage block storage across many nodes. Still, teams often grant the delivery controller broad cluster admin roles during setup. Broad roles let anyone with repo write rights change storage volumes or delete data.
How a longhorn argocd workflow operates
This workflow links your version control system to your cluster storage engine. First, engineers define storage classes, volume claims, and backup targets as declarative files. They push these files to a central Git repo. The GitOps controller watches that repo for any new commits. When a change appears, the controller compares the repo state to the live cluster. Next, it applies the updated manifests to the Kubernetes API server. The storage engine reads those API objects and builds block volumes across host disks. It also sets up volume replicas on separate nodes to protect against disk failure. Every storage change starts as a pull request in Git. Teams review the file changes before the controller runs them in the cluster.
Sync phases and storage engine updates
Block storage engines run many internal components inside a dedicated cluster namespace. They deploy manager pods, driver sets, and user interface services. When the delivery controller syncs all these parts at once, race conditions can happen. A volume claim might apply before the storage driver finishes starting up. To fix this, teams split the setup into ordered sync waves. The first wave creates the namespace and custom resource definitions. The second wave starts the core storage manager and node drivers. The final wave applies recurring snapshot schedules and storage classes. Teams also tell the GitOps tool to ignore live status fields on volume objects. The storage engine updates volume health states every few seconds. Ignoring these runtime status fields stops the controller from showing false out-of-sync alerts. It also prevents endless sync loops during normal disk operations.
What native roles solve and what they miss
Native role-based rules in Kubernetes and GitOps tools control initial login and sync rights. You can map team groups to specific project roles inside the delivery controller. You can also limit which namespaces the controller can write to. These rules decide who can trigger a sync or view the dashboard. Still, static roles do not inspect individual commands or API calls in flight. During a storage outage, engineers often need direct access to the cluster API. They also open the storage engine dashboard to fix stuck volumes or detached disks. Once an engineer logs in with an admin role, every action is open. Native roles cannot block a single destructive delete call while allowing safe read checks. They also fail to mask sensitive data like emails or phone numbers inside logs. Setup rules only check who logs in at the start of a session. They do not govern what individual requests do once the session begins.
ArgoCD Security: architecture patterns and best practices
Continue reading? Get the full guide.
Free. No spam. Unsubscribe anytime.
Add request guardrails and response masking beside Kubernetes
Cluster roles and service accounts govern who reaches the Kubernetes API. Still, a broad role or agent can delete workloads or print sensitive values from ConfigMaps and logs. In practice, hoop.dev runs beside the API server as a sidecar for your Kubernetes API and checks every request and response alongside RBAC.
With hoop.dev Kubernetes controls, guardrails can allow listing Secret names while refusing reads of Secret data before execution. Data masking redacts matched keys in returned JSON responses and log streams in memory. Furthermore, hoop.dev holds SOC 2 Type II and is MIT licensed on GitHub.
Frequently asked questions
Why does my storage app stay stuck in a progressing state during sync?
Storage drivers take time to register with the Kubernetes node agent. If volume claims sync before the driver is ready, the sync waits and stalls.
How do I stop the controller from deleting live volumes during a repo cleanup?
Set a retention policy on your storage class and disable auto-pruning on volume resources. This keeps the physical disk data safe even if someone removes the manifest file.
Can I manage recurring volume backups through Git manifests?
Yes, the storage engine uses custom resources for backup targets and snapshot schedules. You can commit these schedule files to Git just like standard Kubernetes objects.
Related guides
In this post