Skip to content

README.md

Latest commit

 

History

History
118 lines (85 loc) · 3.71 KB

README.md

File metadata and controls

118 lines (85 loc) · 3.71 KB
Raw

agentic-hosting Scripts

Bash automation scripts for common agentic-hosting operations. All scripts read configuration from environment variables.

Setup

export AH_URL="https://<your-domain>"   # ah API base URL
export AH_KEY="keyid.secret"             # Your API key

For local dev:

export AH_URL="http://localhost:8080"

Scripts

Script Description
register.sh Register a new tenant and save credentials
deploy.sh Deploy a service from git URL or Docker image
status.sh Show status of all services and databases
logs.sh Stream build logs for a service
db-provision.sh Provision a database and wire it to a service
health-check.sh Cron-friendly health check with webhook alerting

Usage

chmod +x scripts/*.sh

# Register (one-time, needs bootstrap token)
AH_BOOTSTRAP_TOKEN=<token> ./scripts/register.sh my-tenant me@example.com

# Deploy from git
./scripts/deploy.sh https://github.com/org/repo my-app 3000

# Deploy from Docker image
./scripts/deploy.sh nginx:alpine my-site 80

# Check status
./scripts/status.sh

# Stream build logs
./scripts/logs.sh my-app

# Provision Postgres and wire to service
./scripts/db-provision.sh my-app postgres

# Provision Redis and wire to service
./scripts/db-provision.sh my-app redis

Health Check Script

health-check.sh performs five checks and exits 0 on success or 1 on failure. On failure it POSTs a machine-readable JSON payload to the configured webhook URL.

Checks performed

  1. Health API -- GET /v1/system/health returns {"status":"ok"}
  2. systemd service -- systemctl is-active ah shows active
  3. Infrastructure containers -- docker ps contains paas-traefik and paas-registry
  4. Disk usage -- checks both the ah state dir and Docker data dir, warns at 80%, fails at 90% (configurable)
  5. Public HTTPS -- base domain returns 2xx/3xx (if AH_BASE_DOMAIN is set)

Environment variables

Variable Required Default Description
ALERT_WEBHOOK_URL Yes (unless --dry-run) -- POST endpoint for failure alerts
AH_API_PORT No 8080 Local health API port
AH_SERVICE_NAME No ah systemd service name
AH_BASE_DOMAIN No -- Public domain to probe via HTTPS
AH_DATA_DIR No /var/lib/ah ah state data dir to monitor for disk usage
AH_DOCKER_DATA_DIR No /var/lib/docker Docker data dir to monitor for disk usage
AH_DISK_WARN_PCT No 80 Disk usage warning threshold (%)
AH_DISK_FAIL_PCT No 90 Disk usage failure threshold (%)

Cron setup

Run every 5 minutes (adjust to taste):

*/5 * * * * ALERT_WEBHOOK_URL="https://hooks.example.com/alert" AH_BASE_DOMAIN="agentic.hosting" /opt/agentic-hosting/scripts/health-check.sh >> /var/log/ah-health.log 2>&1

Or with all options:

*/5 * * * * ALERT_WEBHOOK_URL="https://hooks.example.com/alert" AH_API_PORT=8080 AH_SERVICE_NAME=ah AH_BASE_DOMAIN="agentic.hosting" AH_DATA_DIR="/var/lib/ah" AH_DOCKER_DATA_DIR="/var/lib/docker" AH_DISK_WARN_PCT=80 AH_DISK_FAIL_PCT=90 /opt/agentic-hosting/scripts/health-check.sh >> /var/log/ah-health.log 2>&1

Dry-run mode

Test locally without firing the webhook:

./scripts/health-check.sh --dry-run

This runs all checks and prints what would be sent to the webhook, but does not actually POST.

Webhook payload

On failure the script sends a JSON POST body:

{
  "event": "health_check_failed",
  "hostname": "prod-01",
  "timestamp": "2026-03-20T12:00:00Z",
  "failure_count": 1,
  "warning_count": 0,
  "failures": ["Health API returned HTTP 000 (expected 200)"],
  "warnings": []
}