Quick Start

Zero-trust offline diagnostics in 30 seconds. In-cluster Helm install when you're ready.

Srenix runs as a single binary with zero external dependencies. The fastest path is the zero-trust offline mode — no cluster access required, no Helm, no RBAC. When you're ready to watch your cluster continuously, install in-cluster with Helm in three commands.

Prerequisites

Requirement	Minimum	Notes
Kubernetes	1.27	Any distro — EKS, GKE, AKS, k3s, RKE2, OpenShift
kubectl	any	In PATH; kubeconfig pointing at target cluster
helm	3.x	For in-cluster install only
Container pull access	—	Cluster must reach docker.io

Zero-trust offline mode requires none of the above — just the srenix binary and a kubectl get -o json snapshot.

1. Download the binary

Pre-built binaries (release v0.1.0-alpha.1) for linux/amd64, linux/arm64, darwin/amd64, darwin/arm64.

# Linux amd64
curl -sSL https://github.com/srenix-ai/agentic-sre/releases/download/v0.1.0-alpha.1/agentic-sre_linux_amd64.tar.gz \
  | tar xz && sudo mv srenix /usr/local/bin/

# macOS arm64 (Apple Silicon)
curl -sSL https://github.com/srenix-ai/agentic-sre/releases/download/v0.1.0-alpha.1/agentic-sre_darwin_arm64.tar.gz \
  | tar xz && sudo mv srenix /usr/local/bin/

# Other platforms: swap the suffix — linux_arm64 / darwin_amd64

# Verify
srenix version

Newer versions may be available on the releases page — pick a vX.Y.Z release that lists .tar.gz binary assets (chart-only releases don't carry binaries).

2. Zero-trust offline mode (no install needed)

Run diagnostics against a captured snapshot — no cluster RBAC, no write permissions, no Helm chart. Useful for evaluating Srenix before any install.

# Clone the repo to get sample fixtures
git clone https://github.com/srenix-ai/agentic-sre.git
cd agentic-sre

# Run against the bundled sample cluster
srenix diagnose --snapshot examples/sample-cluster

Against a live cluster:

# Capture a snapshot (read-only — never modifies cluster state)
srenix snapshot capture --out ./my-cluster

# Diagnose offline
srenix diagnose --snapshot ./my-cluster

# Or diagnose live directly
srenix diagnose --live

3. In-cluster install via Helm

Installs a watcher Deployment and diagnose/remediate CronJobs. Watches Kubernetes events with a ~10s debounce. The current chart is a SemVer pre-release (0.1.0-alpha.1), so pass --devel (or pin --version 0.1.0-alpha.1) — Helm hides pre-releases otherwise.

helm repo add srenix https://srenix-ai.github.io/agentic-sre
helm repo update
helm install srenix srenix/agentic-sre --devel \
  --namespace agentic-sre --create-namespace \
  --set watcher.enabled=true
# pinned alternative: --version 0.1.0-alpha.1 (instead of --devel)

4. Verify

kubectl get driftreports     # cluster-scoped — no NAMESPACE column
# NAME                       SEVERITY  SOURCE                    SUBJECT                                  LAST SEEN
# drift-4f9d2b8c3e1a7f05     warning   StuckCertificateRequests  CertificateRequest/kube-system/api-tls   12s

kubectl get pods -n agentic-sre

See the DriftReport CRD docs for the full schema and kubectl query patterns.

← Back to docs