mise.toml

[tools]
bitwarden = "latest"
bitwarden-secrets-manager = "2.1.0"
flux2 = "2.9.6"
kubectl = "1.37.1"
kubecolor = "0.8.0"
oxker = "0.13.4"
talosctl = "1.14.5"
usage = "7.1.3"
viddy = "1.3.1"
weave-gitops = "0.38.0"
codex = "0.149.0"
aws-cli = "2.38.6"
k9s = "0.51.0"
java = "27.0.0"

[settings]
auto_install = true

[shell_alias]
ll = "ls -la"
la = "ls -A"
gs = "git status"
gc = "git commit"
k = "kubecolor"
h = "helm"
t = "talosctl"
gto = "gitops"
ft = "forgetool"
tcl = "talosctl"
badPods = 'kubecolor get pods -A --field-selector=status.phase!=Running'
badFlux = "flux get kustomizations -A --status-selector ready=false"
stopAll = "mise chart -stop-all"
kRes = "kubectl rollout restart deployment -n"
vK = "viddy kubectl get pods,ingress,configmaps"
getPods = "mise chart -pods-info -n"
sebDC = "docker compose -f docker/.doco-cd/compose.yaml"
dcLogs = "docker logs doco-cd -f"
kHighRes = "kubectl get pods -A --no-headers | awk '$5 > 0' | sort -k5,5nr"
ingress = "mise chart -ingress"
dns = "mise chart -dns"
deFrag = "mise talos:etcd:status --defrag --yes"
delPods = "mise chart --pods-delete --confirm"
podInfo = "mise chart -pods-info"
volClean = "mise volsync --cleanup --dry-run"
kUps = "mise run kUps --"
cnpgRepair = "mise run cnpg:repair --"

[tasks.kUps]
description = "Get or describe Kubernetes and Talos upgrades"
run = '''
#!/usr/bin/env bash
set -euo pipefail

case "${1:-}" in
  -g|--get)
    action="get"
    ;;
  -d|--describe)
    action="describe"
    ;;
  *)
    echo "Usage: kUps (-g|--get|-d|--describe)" >&2
    exit 2
    ;;
esac

kubecolor "$action" kubernetesupgrades,talosupgrades
'''

[tasks."cnpg:repair"]
description = "Diagnose unhealthy CNPG clusters and optionally recycle unhealthy replicas"
run = '''
#!/usr/bin/env bash
set -euo pipefail

namespace=""
apply=false
all_namespaces=true

usage() {
  cat <<'EOF'
Usage: mise run cnpg:repair -- [-n <namespace> | -A] [--apply]

Defaults to all namespaces.

Diagnoses every CloudNativePG Cluster in a namespace or across all namespaces.
With --apply, deletes
only unhealthy replica pods so the CNPG operator can recreate them. It never
deletes a current primary or any PVC.

Examples:
  mise run cnpg:repair
  mise run cnpg:repair -- -n ggrequestz
  mise run cnpg:repair -- -A
  mise run cnpg:repair -- -n ggrequestz --apply
EOF
}

while (($#)); do
  case "$1" in
    -n|--namespace)
      namespace="${2:-}"
      all_namespaces=false
      shift 2
      ;;
    -A|--all-namespaces)
      all_namespaces=true
      shift
      ;;
    --apply)
      apply=true
      shift
      ;;
    -h|--help)
      usage
      exit 0
      ;;
    *)
      echo "Unknown option: $1" >&2
      usage >&2
      exit 2
      ;;
  esac
done

if [[ "$all_namespaces" == true && -n "$namespace" ]]; then
  echo "Use either --namespace or --all-namespaces, not both." >&2
  usage >&2
  exit 2
fi

if [[ "$all_namespaces" == true ]]; then
  mapfile -t clusters < <(kubectl get cluster.postgresql.cnpg.io -A -o custom-columns=NAMESPACE:.metadata.namespace,NAME:.metadata.name --no-headers)
else
  mapfile -t clusters < <(kubectl get cluster.postgresql.cnpg.io -n "$namespace" -o custom-columns=NAME:.metadata.name --no-headers)
fi

if [[ -z "${clusters[0]:-}" ]]; then
  if [[ "$all_namespaces" == true ]]; then
    echo "No CNPG clusters found."
  else
    echo "No CNPG clusters found in namespace: $namespace"
  fi
  exit 0
fi

for cluster_ref in "${clusters[@]}"; do
  if [[ "$all_namespaces" == true ]]; then
    read -r cluster_namespace cluster <<< "$cluster_ref"
  else
    cluster_namespace="$namespace"
    cluster="$cluster_ref"
  fi
  primary="$(kubectl get cluster.postgresql.cnpg.io "$cluster" -n "$cluster_namespace" -o jsonpath='{.status.currentPrimary}')"
  ready="$(kubectl get cluster.postgresql.cnpg.io "$cluster" -n "$cluster_namespace" -o jsonpath='{range .status.conditions[?(@.type=="Ready")]}{.status}{end}')"
  archiving="$(kubectl get cluster.postgresql.cnpg.io "$cluster" -n "$cluster_namespace" -o jsonpath='{range .status.conditions[?(@.type=="ContinuousArchiving")]}{.status}{end}')"

  echo
  echo "CNPG cluster: $cluster_namespace/$cluster"
  echo "  primary: ${primary:-unknown}"
  echo "  ready: ${ready:-unknown}; continuous archiving: ${archiving:-not configured}"
  kubectl get pods -n "$cluster_namespace" -l "cnpg.io/cluster=$cluster" \
    -o custom-columns=NAME:.metadata.name,READY:.status.containerStatuses[0].ready,PHASE:.status.phase,RESTARTS:.status.containerStatuses[0].restartCount --no-headers

  mapfile -t instances < <(kubectl get pods -n "$cluster_namespace" -l "cnpg.io/cluster=$cluster" -o custom-columns=NAME:.metadata.name --no-headers)

  for pod in "${instances[@]}"; do
    [[ "$pod" == "$primary" ]] && continue
    pod_ready="$(kubectl get pod "$pod" -n "$cluster_namespace" -o jsonpath='{range .status.conditions[?(@.type=="Ready")]}{.status}{end}')"
    if [[ "$pod_ready" == "True" ]]; then
      continue
    fi

    echo "  unhealthy replica: $pod"
    kubectl logs -n "$cluster_namespace" "$pod" --tail=25 2>/dev/null \
      | grep -E -i 'error|fatal|startup probe|archive|wal|no such bucket' \
      | tail -10 \
      | sed 's/^/    /' || true

    if [[ "$apply" == true ]]; then
      echo "  recycling unhealthy replica: $pod"
      kubectl delete pod -n "$cluster_namespace" "$pod"
    else
      echo "  diagnostic only; rerun with --apply to recycle this replica"
    fi
  done

  if [[ "$archiving" == "False" ]]; then
    echo "  warning: continuous archiving is failing; inspect the log lines above before recycling replicas"
  fi
done
'''