feat(k8s): report instance counters onto the deployment during the wait

CHANGELOG.md

@@ -6,6 +6,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+- k8s scope deployments now report launched and healthy instance counts, so the deployment page shows live "X/Y launched" and "X/Y healthy" progress
 - Fix: triggering a scheduled task job on a scope that is not deployed now fails with a clear "deploy the scope first" message instead of an opaque error
 - Add "Kill instance" action to scheduled task scopes to terminate an individual running job instance
 

k8s/deployment/tests/wait_deployment_active.bats

@@ -703,3 +703,157 @@ teardown() {
     return 1
   fi
 }
+
+# =============================================================================
+# Instance Count Reporting Tests
+# =============================================================================
+@test "wait_deployment_active: reports the instance counters as a partial strategy_data patch" {
+  export PATCH_LOG="$BATS_TEST_TMPDIR/patches.log"
+
+  kubectl() {
+    case "$*" in
+      "get deployment"*"-o json"*)
+        echo '{
+          "spec": {"replicas": 3},
+          "status": {
+            "replicas": 3,
+            "availableReplicas": 3,
+            "updatedReplicas": 3,
+            "readyReplicas": 3
+          }
+        }'
+        ;;
+      "get pods"*)
+        echo ""
+        ;;
+      "get events"*)
+        echo '{"items":[]}'
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  np() {
+    case "$*" in
+      *"deployment patch"*)
+        echo "$*" >> "$PATCH_LOG"
+        ;;
+      *)
+        echo "running"
+        ;;
+    esac
+  }
+  export -f np
+
+  run bash "$BATS_TEST_DIRNAME/../wait_deployment_active"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "📡 Reported instance counts — launched: 3/3, healthy: 3/3"
+  [ "$(wc -l < "$PATCH_LOG")" -eq 1 ]
+  assert_contains "$(cat "$PATCH_LOG")" '"amount_instances_to_wait": 3'
+  assert_contains "$(cat "$PATCH_LOG")" '"launched_instances": 3'
+  assert_contains "$(cat "$PATCH_LOG")" '"healthy_instances": 3'
+  # ONLY the counters travel — the server merges them over the current
+  # strategy_data, so nothing else may ride (or wipe) the body.
+  if grep -q "switched_traffic" "$PATCH_LOG"; then
+    echo "Expected the patch body to carry only the counters"
+    cat "$PATCH_LOG"
+    return 1
+  fi
+}
+
+@test "wait_deployment_active: re-reports only when the counts change" {
+  export PATCH_LOG="$BATS_TEST_TMPDIR/patches.log"
+  export KUBECTL_CALLS="$BATS_TEST_TMPDIR/kubectl.calls"
+
+  run bash -c "
+    sleep() { :; }
+    export -f sleep
+
+    kubectl() {
+      case \"\$*\" in
+        \"get deployment\"*\"-o json\"*)
+          echo x >> \"\$KUBECTL_CALLS\"
+          if [ \"\$(wc -l < \"\$KUBECTL_CALLS\")\" -lt 3 ]; then
+            echo '{\"spec\": {\"replicas\": 3}, \"status\": {\"replicas\": 3, \"availableReplicas\": 1, \"updatedReplicas\": 3, \"readyReplicas\": 1}}'
+          else
+            echo '{\"spec\": {\"replicas\": 3}, \"status\": {\"replicas\": 3, \"availableReplicas\": 3, \"updatedReplicas\": 3, \"readyReplicas\": 3}}'
+          fi
+          ;;
+        \"get pods\"*)
+          echo ''
+          ;;
+        \"get events\"*)
+          echo '{\"items\":[]}'
+          ;;
+      esac
+    }
+    export -f kubectl
+
+    np() {
+      case \"\$*\" in
+        *'deployment patch'*)
+          echo \"\$*\" >> \"\$PATCH_LOG\"
+          ;;
+        *)
+          echo 'running'
+          ;;
+      esac
+    }
+    export -f np
+
+    export SERVICE_PATH='$SERVICE_PATH' K8S_NAMESPACE='$K8S_NAMESPACE'
+    export SCOPE_ID='$SCOPE_ID' DEPLOYMENT_ID='$DEPLOYMENT_ID'
+    export TIMEOUT=30 NP_API_KEY='$NP_API_KEY' SKIP_DEPLOYMENT_STATUS_CHECK='false'
+    bash '$BATS_TEST_DIRNAME/../wait_deployment_active'
+  "
+
+  [ "$status" -eq 0 ]
+  # Three iterations (1/3, 1/3, 3/3) but only two distinct counts — two patches.
+  [ "$(wc -l < "$PATCH_LOG")" -eq 2 ]
+  assert_contains "$(sed -n 1p "$PATCH_LOG")" '"healthy_instances": 1'
+  assert_contains "$(sed -n 2p "$PATCH_LOG")" '"healthy_instances": 3'
+}
+
+@test "wait_deployment_active: a failed count report never fails the deployment" {
+  kubectl() {
+    case "$*" in
+      "get deployment"*"-o json"*)
+        echo '{
+          "spec": {"replicas": 3},
+          "status": {
+            "replicas": 3,
+            "availableReplicas": 3,
+            "updatedReplicas": 3,
+            "readyReplicas": 3
+          }
+        }'
+        ;;
+      "get pods"*)
+        echo ""
+        ;;
+      "get events"*)
+        echo '{"items":[]}'
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  np() {
+    case "$*" in
+      *"deployment patch"*)
+        return 1
+        ;;
+      *)
+        echo "running"
+        ;;
+    esac
+  }
+  export -f np
+
+  run bash "$BATS_TEST_DIRNAME/../wait_deployment_active"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "Could not report instance counts"
+  assert_contains "$output" "✅ All pods in deployment 'd-scope-123-deploy-456' are available and ready!"
+}

k8s/deployment/wait_deployment_active

@@ -54,6 +54,38 @@ K8S_DEPLOYMENT_NAME="d-$SCOPE_ID-$DEPLOYMENT_ID"
 iteration=0
 LATEST_TIMESTAMP=$(date -u +"%Y-%m-%dT%H:%M:%SZ")
 SKIP_DEPLOYMENT_STATUS_CHECK="${SKIP_DEPLOYMENT_STATUS_CHECK:=false}"
+LAST_REPORTED_COUNTS=""
+
+# Report the instance counters onto the deployment's strategy_data
+# (amount_instances_to_wait / launched_instances / healthy_instances) so the
+# platform can show launched/healthy progress. A partial strategy_data body
+# MERGES server-side (the deployment strategies fold the update over the
+# current data), so only the counters travel — switched_traffic and friends are
+# untouched. Same contract the scope-workflow-manager uses for its scopes.
+# Best-effort telemetry: a failed report must never fail the deployment.
+report_instance_counts() {
+    local desired="$1" launched="$2" healthy="$3"
+
+    if ! [[ "$desired" =~ ^[0-9]+$ ]] || [ "$desired" -eq 0 ]; then
+        return 0
+    fi
+    [[ "$launched" =~ ^[0-9]+$ ]] || launched=0
+    [[ "$healthy" =~ ^[0-9]+$ ]] || healthy=0
+
+    local counts="$desired/$launched/$healthy"
+    if [ "$counts" = "$LAST_REPORTED_COUNTS" ]; then
+        return 0
+    fi
+
+    local body="{\"strategy_data\": {\"amount_instances_to_wait\": $desired, \"launched_instances\": $launched, \"healthy_instances\": $healthy}}"
+    if np deployment patch --id "$DEPLOYMENT_ID" --api-key "$NP_API_KEY" --no-output --body "$body" 2>/dev/null; then
+        LAST_REPORTED_COUNTS="$counts"
+        log debug "📡 Reported instance counts — launched: $launched/$desired, healthy: $healthy/$desired"
+    else
+        log debug "Could not report instance counts (will retry on next change)"
+    fi
+    return 0
+}
 
 HEARTBEAT_INTERVAL=$(( MAX_ITERATIONS / 10 ))
 [ "$HEARTBEAT_INTERVAL" -lt 1 ] && HEARTBEAT_INTERVAL=1
@@ -104,8 +136,11 @@ while true; do
     current=$(echo "$deployment_status" | jq '.status.availableReplicas // 0')
     updated=$(echo "$deployment_status" | jq '.status.updatedReplicas // 0')
     ready=$(echo "$deployment_status" | jq '.status.readyReplicas // 0')
+    launched=$(echo "$deployment_status" | jq '.status.replicas // 0')
     log debug "🔍 $(date): Iteration $iteration - Deployment status - Available: $current/$desired, Updated: $updated/$desired, Ready: $ready/$desired"
 
+    report_instance_counts "$desired" "$launched" "$ready"
+
     if [ "$desired" = "$current" ] && [ "$desired" = "$updated" ] && [ "$desired" = "$ready" ] && [ "$desired" -gt 0 ]; then
         log debug ""
         log info "✅ All pods in deployment '$K8S_DEPLOYMENT_NAME' are available and ready!"