From b22d76310509e616942a0d39b116fb08c7858ad9 Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 13:00:23 -0300
Subject: [PATCH 1/7] fix: clear error when triggering job on undeployed
 scheduled task

---
 CHANGELOG.md                            |   3 +
 scheduled_task/scope/tests/trigger.bats | 174 ++++++++++++++++++++++++
 scheduled_task/scope/trigger            |  31 ++++-
 3 files changed, 207 insertions(+), 1 deletion(-)
 create mode 100644 scheduled_task/scope/tests/trigger.bats

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 1aa174b9..85362f20 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,9 @@ All notable changes to this project will be documented in this file.
 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]
+- 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
+
 ## [1.13.0] - 2026-07-10
 - Add support to get AWS credentials via assume role
 - Add support to auto-create ALBs on scope create
diff --git a/scheduled_task/scope/tests/trigger.bats b/scheduled_task/scope/tests/trigger.bats
new file mode 100644
index 00000000..b7cd1d75
--- /dev/null
+++ b/scheduled_task/scope/tests/trigger.bats
@@ -0,0 +1,174 @@
+#!/usr/bin/env bats
+# =============================================================================
+# Unit tests for scheduled_task scope/trigger - manually trigger the CronJob
+#
+# Contract:
+#   - A scope with no active deployment has no CronJob, so triggering must fail
+#     early with a clear "deploy first" message instead of an opaque kubectl
+#     error from `kubectl create job --from=cronjob/`.
+# =============================================================================
+
+setup() {
+  export PROJECT_ROOT="$(cd "$BATS_TEST_DIRNAME/../../.." && pwd)"
+  source "$PROJECT_ROOT/testing/assertions.sh"
+
+  export K8S_NAMESPACE="default-namespace"
+
+  # Base CONTEXT: deployed scope with an active deployment
+  export CONTEXT='{
+    "scope": {
+      "id": "scope-123",
+      "current_active_deployment": "deploy-456"
+    },
+    "providers": {
+      "container-orchestration": {
+        "cluster": {
+          "namespace": "provider-namespace"
+        }
+      }
+    }
+  }'
+
+  # Mock kubectl: cronjob exists, job creation succeeds
+  kubectl() {
+    case "$*" in
+      "get cronjob -n provider-namespace -l scope_id=scope-123 -o jsonpath={.items[0].metadata.name}")
+        echo "my-cronjob"
+        return 0
+        ;;
+      "create job --from=cronjob/my-cronjob"*)
+        return 0
+        ;;
+      *)
+        return 0
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  # Deterministic timestamp for job name suffix
+  date() { echo "1700000000"; }
+  export -f date
+}
+
+teardown() {
+  unset -f kubectl date
+}
+
+# =============================================================================
+# Success Flow
+# =============================================================================
+@test "trigger: success flow - finds cronjob and triggers a job" {
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "Triggering job my-cronjob"
+  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+}
+
+# =============================================================================
+# Error: scope not deployed (no active deployment)
+# =============================================================================
+@test "trigger: fails with a clear message when the scope has no active deployment" {
+  export CONTEXT=$(echo "$CONTEXT" | jq 'del(.scope.current_active_deployment)')
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ The scope has no active deployment, so there is no scheduled job to trigger"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "The scope has not been deployed yet"
+  assert_contains "$output" "🔧 How to fix:"
+  assert_contains "$output" "• Deploy the scope first, then trigger the job"
+}
+
+@test "trigger: fails when current_active_deployment is null" {
+  export CONTEXT=$(echo "$CONTEXT" | jq '.scope.current_active_deployment = null')
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ The scope has no active deployment, so there is no scheduled job to trigger"
+  assert_contains "$output" "• Deploy the scope first, then trigger the job"
+}
+
+# =============================================================================
+# Error: deployed but no CronJob found
+# =============================================================================
+@test "trigger: fails with a clear message when no CronJob exists for the scope" {
+  kubectl() {
+    case "$*" in
+      "get cronjob -n provider-namespace -l scope_id=scope-123 -o jsonpath={.items[0].metadata.name}")
+        echo ""
+        return 0
+        ;;
+      *)
+        return 0
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ No CronJob found for scope 'scope-123' in namespace 'provider-namespace'"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "🔧 How to fix:"
+}
+
+# =============================================================================
+# Namespace resolution
+# =============================================================================
+@test "trigger: uses the namespace from the provider for lookup and job creation" {
+  # kubectl only succeeds when called against the provider namespace; any other
+  # namespace makes it fail, so this asserts the resolved namespace is threaded
+  # through both the cronjob lookup and the job creation.
+  kubectl() {
+    case "$*" in
+      "get cronjob -n provider-namespace"*)
+        echo "my-cronjob"
+        return 0
+        ;;
+      "create job --from=cronjob/my-cronjob"*"-n provider-namespace")
+        return 0
+        ;;
+      *)
+        echo "unexpected namespace: $*" >&2
+        return 1
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+}
+
+@test "trigger: falls back to the default namespace when the provider namespace is not set" {
+  export CONTEXT=$(echo "$CONTEXT" | jq 'del(.providers["container-orchestration"].cluster.namespace)')
+
+  kubectl() {
+    case "$*" in
+      "get cronjob -n default-namespace"*)
+        echo "my-cronjob"
+        return 0
+        ;;
+      "create job --from=cronjob/my-cronjob"*"-n default-namespace")
+        return 0
+        ;;
+      *)
+        echo "unexpected namespace: $*" >&2
+        return 1
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+}
diff --git a/scheduled_task/scope/trigger b/scheduled_task/scope/trigger
index 1ebf6cce..ae1f2928 100644
--- a/scheduled_task/scope/trigger
+++ b/scheduled_task/scope/trigger
@@ -1,5 +1,21 @@
 #!/bin/bash
 
+set -euo pipefail
+
+DEPLOYMENT_ID=$(echo "$CONTEXT" | jq -r '.scope.current_active_deployment // empty')
+
+if [[ -z "$DEPLOYMENT_ID" ]]; then
+  echo "❌ The scope has no active deployment, so there is no scheduled job to trigger" >&2
+  echo "" >&2
+  echo "💡 Possible causes:" >&2
+  echo "   The scope has not been deployed yet" >&2
+  echo "" >&2
+  echo "🔧 How to fix:" >&2
+  echo "   • Deploy the scope first, then trigger the job" >&2
+  echo "" >&2
+  exit 1
+fi
+
 K8S_NAMESPACE=$(echo "$CONTEXT" | jq -r --arg default "$K8S_NAMESPACE" '
   .providers["container-orchestration"].cluster.namespace // $default
 ')
@@ -8,8 +24,21 @@ SCOPE_ID=$(echo "$CONTEXT" | jq -r '.scope.id')
 
 JOB=$(kubectl get cronjob -n "$K8S_NAMESPACE" -l "scope_id=$SCOPE_ID" -o jsonpath="{.items[0].metadata.name}" 2>/dev/null || echo "")
 
+if [[ -z "$JOB" ]]; then
+  echo "❌ No CronJob found for scope '$SCOPE_ID' in namespace '$K8S_NAMESPACE'" >&2
+  echo "" >&2
+  echo "💡 Possible causes:" >&2
+  echo "   The scope's scheduled job may not have been created yet, or the deployment is still in progress" >&2
+  echo "" >&2
+  echo "🔧 How to fix:" >&2
+  echo "   • Verify the CronJob exists: kubectl get cronjob -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID" >&2
+  echo "   • Redeploy the scope if the CronJob is missing" >&2
+  echo "" >&2
+  exit 1
+fi
+
 echo "Triggering job $JOB"
 
 kubectl create job --from="cronjob/$JOB" "$JOB-$(date +%s)" -n "$K8S_NAMESPACE"
 
-echo "The job $JOB was triggered, you can follow the execution from the logs screen"
\ No newline at end of file
+echo "The job $JOB was triggered, you can follow the execution from the logs screen"

From d4b389504df42a57f0df439408ccbbb191eaa9c4 Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 13:54:16 -0300
Subject: [PATCH 2/7] refactor: use log helper in scheduled task trigger via
 workflow step

---
 scheduled_task/logging                        | 41 +++++++++++++++++++
 scheduled_task/scope/tests/trigger.bats       | 15 ++++---
 scheduled_task/scope/trigger                  | 38 ++++++++---------
 .../scope/workflows/trigger-job.yaml          | 11 ++++-
 4 files changed, 80 insertions(+), 25 deletions(-)
 create mode 100644 scheduled_task/logging

diff --git a/scheduled_task/logging b/scheduled_task/logging
new file mode 100644
index 00000000..d0df55d7
--- /dev/null
+++ b/scheduled_task/logging
@@ -0,0 +1,41 @@
+#!/bin/bash
+
+# Logging utility — log4j-style level filtering
+# Usage:    log "level" "message"
+# Levels:   debug < info < warn < error
+# Control:  LOG_LEVEL env var (default: info)
+#
+# Example:
+#   LOG_LEVEL=info
+#   log debug "verbose details"  # suppressed
+#   log info  "deployment done"  # shown
+log() {
+  local level="${1:-info}"
+  local message="${2:-}"
+
+  local -i msg_num threshold
+
+  case "${level,,}" in
+    debug) msg_num=0 ;;
+    info)  msg_num=1 ;;
+    warn)  msg_num=2 ;;
+    error) msg_num=3 ;;
+    *)     msg_num=1 ;;
+  esac
+
+  case "${LOG_LEVEL:-info}" in
+    debug) threshold=0 ;;
+    info)  threshold=1 ;;
+    warn)  threshold=2 ;;
+    error) threshold=3 ;;
+    *)     threshold=1 ;;
+  esac
+
+  if [ "$msg_num" -ge "$threshold" ]; then
+    if [ "$msg_num" -ge 3 ]; then
+      echo "$message" >&2
+    else
+      echo "$message"
+    fi
+  fi
+}
diff --git a/scheduled_task/scope/tests/trigger.bats b/scheduled_task/scope/tests/trigger.bats
index b7cd1d75..92e33fad 100644
--- a/scheduled_task/scope/tests/trigger.bats
+++ b/scheduled_task/scope/tests/trigger.bats
@@ -12,6 +12,11 @@ setup() {
   export PROJECT_ROOT="$(cd "$BATS_TEST_DIRNAME/../../.." && pwd)"
   source "$PROJECT_ROOT/testing/assertions.sh"
 
+  # The workflow loads the `log` function via a `load logging` step, so the
+  # script assumes it exists. Mock it here (errors go to stderr, like the real one).
+  log() { if [ "$1" = "error" ]; then echo "$2" >&2; else echo "$2"; fi; }
+  export -f log
+
   export K8S_NAMESPACE="default-namespace"
 
   # Base CONTEXT: deployed scope with an active deployment
@@ -52,7 +57,7 @@ setup() {
 }
 
 teardown() {
-  unset -f kubectl date
+  unset -f kubectl date log
 }
 
 # =============================================================================
@@ -62,8 +67,8 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../trigger"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "Triggering job my-cronjob"
-  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+  assert_contains "$output" "📝 Triggering job my-cronjob"
+  assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
 
 # =============================================================================
@@ -144,7 +149,7 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../trigger"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+  assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
 
 @test "trigger: falls back to the default namespace when the provider namespace is not set" {
@@ -170,5 +175,5 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../trigger"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "The job my-cronjob was triggered, you can follow the execution from the logs screen"
+  assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
diff --git a/scheduled_task/scope/trigger b/scheduled_task/scope/trigger
index ae1f2928..7a2f47d7 100644
--- a/scheduled_task/scope/trigger
+++ b/scheduled_task/scope/trigger
@@ -5,14 +5,14 @@ set -euo pipefail
 DEPLOYMENT_ID=$(echo "$CONTEXT" | jq -r '.scope.current_active_deployment // empty')
 
 if [[ -z "$DEPLOYMENT_ID" ]]; then
-  echo "❌ The scope has no active deployment, so there is no scheduled job to trigger" >&2
-  echo "" >&2
-  echo "💡 Possible causes:" >&2
-  echo "   The scope has not been deployed yet" >&2
-  echo "" >&2
-  echo "🔧 How to fix:" >&2
-  echo "   • Deploy the scope first, then trigger the job" >&2
-  echo "" >&2
+  log error "❌ The scope has no active deployment, so there is no scheduled job to trigger"
+  log error ""
+  log error "💡 Possible causes:"
+  log error "   The scope has not been deployed yet"
+  log error ""
+  log error "🔧 How to fix:"
+  log error "   • Deploy the scope first, then trigger the job"
+  log error ""
   exit 1
 fi
 
@@ -25,20 +25,20 @@ SCOPE_ID=$(echo "$CONTEXT" | jq -r '.scope.id')
 JOB=$(kubectl get cronjob -n "$K8S_NAMESPACE" -l "scope_id=$SCOPE_ID" -o jsonpath="{.items[0].metadata.name}" 2>/dev/null || echo "")
 
 if [[ -z "$JOB" ]]; then
-  echo "❌ No CronJob found for scope '$SCOPE_ID' in namespace '$K8S_NAMESPACE'" >&2
-  echo "" >&2
-  echo "💡 Possible causes:" >&2
-  echo "   The scope's scheduled job may not have been created yet, or the deployment is still in progress" >&2
-  echo "" >&2
-  echo "🔧 How to fix:" >&2
-  echo "   • Verify the CronJob exists: kubectl get cronjob -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID" >&2
-  echo "   • Redeploy the scope if the CronJob is missing" >&2
-  echo "" >&2
+  log error "❌ No CronJob found for scope '$SCOPE_ID' in namespace '$K8S_NAMESPACE'"
+  log error ""
+  log error "💡 Possible causes:"
+  log error "   The scope's scheduled job may not have been created yet, or the deployment is still in progress"
+  log error ""
+  log error "🔧 How to fix:"
+  log error "   • Verify the CronJob exists: kubectl get cronjob -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID"
+  log error "   • Redeploy the scope if the CronJob is missing"
+  log error ""
   exit 1
 fi
 
-echo "Triggering job $JOB"
+log info "📝 Triggering job $JOB"
 
 kubectl create job --from="cronjob/$JOB" "$JOB-$(date +%s)" -n "$K8S_NAMESPACE"
 
-echo "The job $JOB was triggered, you can follow the execution from the logs screen"
+log info "✅ The job $JOB was triggered, you can follow the execution from the logs screen"
diff --git a/scheduled_task/scope/workflows/trigger-job.yaml b/scheduled_task/scope/workflows/trigger-job.yaml
index 6b07c1ce..954bccbd 100644
--- a/scheduled_task/scope/workflows/trigger-job.yaml
+++ b/scheduled_task/scope/workflows/trigger-job.yaml
@@ -2,6 +2,15 @@ provider_categories:
   - container-orchestration
   - cloud-providers
 steps:
+  - name: load logging
+    type: script
+    file: "$OVERRIDES_PATH/logging"
+    output:
+      - name: log
+        type: function
+        parameters:
+          level: string
+          message: string
   - name: trigger
     type: script
-    file: "$OVERRIDES_PATH/scope/trigger"
\ No newline at end of file
+    file: "$OVERRIDES_PATH/scope/trigger"

From 9e5581a57d11139cfec5cb4bb0654a9e40956223 Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 13:54:27 -0300
Subject: [PATCH 3/7] feat: add kill-instances action to scheduled task scope

---
 CHANGELOG.md                                  |   1 +
 scheduled_task/deployment/kill_instances      | 126 ++++++
 .../deployment/tests/kill_instances.bats      | 412 ++++++++++++++++++
 .../deployment/tests/workflow_overrides.bats  |  40 ++
 .../deployment/workflows/kill_instances.yaml  |  17 +
 .../specs/actions/kill-instances.json.tpl     |  18 +
 6 files changed, 614 insertions(+)
 create mode 100755 scheduled_task/deployment/kill_instances
 create mode 100644 scheduled_task/deployment/tests/kill_instances.bats
 create mode 100644 scheduled_task/deployment/workflows/kill_instances.yaml
 create mode 100644 scheduled_task/specs/actions/kill-instances.json.tpl

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 85362f20..9a7b51bd 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 - 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 instances" action to scheduled task scopes to terminate an individual running job instance
 
 ## [1.13.0] - 2026-07-10
 - Add support to get AWS credentials via assume role
diff --git a/scheduled_task/deployment/kill_instances b/scheduled_task/deployment/kill_instances
new file mode 100755
index 00000000..4a0bb22d
--- /dev/null
+++ b/scheduled_task/deployment/kill_instances
@@ -0,0 +1,126 @@
+#!/bin/bash
+
+set -euo pipefail
+
+
+log debug "🔍 Starting instance kill operation..."
+
+DEPLOYMENT_ID=$(echo "$CONTEXT" | jq -r '.parameters.deployment_id // .notification.parameters.deployment_id // empty')
+INSTANCE_NAME=$(echo "$CONTEXT" | jq -r '.parameters.instance_name // .notification.parameters.instance_name // empty')
+
+if [[ -z "$DEPLOYMENT_ID" ]] && [[ -n "${NP_ACTION_CONTEXT:-}" ]]; then
+    DEPLOYMENT_ID=$(echo "$NP_ACTION_CONTEXT" | jq -r '.notification.parameters.deployment_id // empty')
+fi
+
+if [[ -z "$INSTANCE_NAME" ]] && [[ -n "${NP_ACTION_CONTEXT:-}" ]]; then
+    INSTANCE_NAME=$(echo "$NP_ACTION_CONTEXT" | jq -r '.notification.parameters.instance_name // empty')
+fi
+
+if [[ -z "$DEPLOYMENT_ID" ]]; then
+    log error "❌ deployment_id parameter not found"
+    log error "💡 Possible causes:"
+    log error "   - Parameter not provided in action request"
+    log error "   - Context structure is different than expected"
+    log error "🔧 How to fix:"
+    log error "   - Ensure deployment_id is passed in the action parameters"
+    exit 1
+fi
+
+if [[ -z "$INSTANCE_NAME" ]]; then
+    log error "❌ instance_name parameter not found"
+    log error "💡 Possible causes:"
+    log error "   - Parameter not provided in action request"
+    log error "   - Context structure is different than expected"
+    log error "🔧 How to fix:"
+    log error "   - Ensure instance_name is passed in the action parameters"
+    exit 1
+fi
+
+log debug "📋 Deployment ID: $DEPLOYMENT_ID"
+log debug "📋 Instance name: $INSTANCE_NAME"
+
+SCOPE_ID=$(echo "$CONTEXT" | jq -r '.tags.scope_id // .scope.id // .notification.tags.scope_id // empty')
+
+if [[ -z "$SCOPE_ID" ]] && [[ -n "${NP_ACTION_CONTEXT:-}" ]]; then
+    SCOPE_ID=$(echo "$NP_ACTION_CONTEXT" | jq -r '.notification.tags.scope_id // empty')
+fi
+
+K8S_NAMESPACE=$(echo "$CONTEXT" | jq -r --arg default "$K8S_NAMESPACE" '
+  .providers["container-orchestration"].cluster.namespace // $default
+' 2>/dev/null || echo "nullplatform")
+
+if [[ -z "$SCOPE_ID" ]]; then
+    log error "❌ scope_id not found in context"
+    log error "💡 Possible causes:"
+    log error "   - Context missing scope information"
+    log error "   - Action invoked outside of scope context"
+    log error "🔧 How to fix:"
+    log error "   - Verify the action is invoked with proper scope context"
+    exit 1
+fi
+
+log debug "📋 Scope ID: $SCOPE_ID"
+log debug "📋 Namespace: $K8S_NAMESPACE"
+
+log debug "🔍 Verifying pod exists..."
+if ! kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
+    log error "❌ Pod $INSTANCE_NAME not found in namespace $K8S_NAMESPACE"
+    log error "💡 Possible causes:"
+    log error "   - Pod was already terminated"
+    log error "   - Pod name is incorrect"
+    log error "   - Pod exists in a different namespace"
+    log error "🔧 How to fix:"
+    log error "   - List pods: kubectl get pods -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID"
+    exit 1
+fi
+
+log debug "📋 Fetching pod details..."
+POD_STATUS=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.phase}')
+POD_NODE=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.spec.nodeName}')
+POD_START_TIME=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.startTime}')
+
+log debug "📋 Pod: $INSTANCE_NAME | Status: $POD_STATUS | Node: $POD_NODE | Started: $POD_START_TIME"
+
+CRONJOB_NAME="job-$SCOPE_ID-$DEPLOYMENT_ID"
+
+POD_JOB=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.metadata.ownerReferences[0].name}' 2>/dev/null || echo "")
+if [[ -n "$POD_JOB" ]]; then
+    JOB_CRONJOB=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.metadata.ownerReferences[0].name}' 2>/dev/null || echo "")
+    log debug "📋 Pod ownership: Job=$POD_JOB -> CronJob=$JOB_CRONJOB"
+
+    if [[ "$JOB_CRONJOB" != "$CRONJOB_NAME" ]]; then
+        log warn "⚠️  Pod does not belong to expected scheduled task $CRONJOB_NAME (continuing anyway)"
+    fi
+else
+    log warn "⚠️  Could not verify pod ownership"
+fi
+
+log debug "📝 Deleting pod $INSTANCE_NAME with 30s grace period..."
+kubectl delete pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" --grace-period=30
+
+log debug "📝 Waiting for pod termination..."
+kubectl wait --for=delete pod/"$INSTANCE_NAME" -n "$K8S_NAMESPACE" --timeout=60s || log warn "⚠️  Pod deletion timeout reached"
+
+if kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
+    POD_STATUS_AFTER=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.phase}')
+    log warn "⚠️  Pod still exists after deletion attempt (status: $POD_STATUS_AFTER)"
+else
+    log info "✅ Pod successfully terminated and removed"
+fi
+
+log debug "📋 Checking job status after pod deletion..."
+if [[ -n "$POD_JOB" ]] && kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
+    JOB_ACTIVE=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.status.active}')
+    JOB_SUCCEEDED=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.status.succeeded}')
+    JOB_FAILED=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.status.failed}')
+
+    log debug "📋 Job $POD_JOB: active=${JOB_ACTIVE:-0}, succeeded=${JOB_SUCCEEDED:-0}, failed=${JOB_FAILED:-0}"
+
+    if [[ "${JOB_ACTIVE:-0}" -gt 0 ]]; then
+        log debug "📋 Job is still active; Kubernetes may start a replacement pod (backoffLimit permitting)"
+    fi
+else
+    log warn "⚠️  Job for pod not found (it may have already completed)"
+fi
+
+log info "✨ Instance kill operation completed for $INSTANCE_NAME"
diff --git a/scheduled_task/deployment/tests/kill_instances.bats b/scheduled_task/deployment/tests/kill_instances.bats
new file mode 100644
index 00000000..6f1d86e4
--- /dev/null
+++ b/scheduled_task/deployment/tests/kill_instances.bats
@@ -0,0 +1,412 @@
+#!/usr/bin/env bats
+# =============================================================================
+# Unit tests for scheduled_task deployment/kill_instances - job pod termination
+#
+# scheduled_task pods are owned by Job -> CronJob (job-<scope>-<deployment>),
+# unlike the base k8s scope where pods are owned by ReplicaSet -> Deployment.
+# =============================================================================
+
+setup() {
+  export PROJECT_ROOT="$(cd "$BATS_TEST_DIRNAME/../../.." && pwd)"
+  source "$PROJECT_ROOT/testing/assertions.sh"
+  log() { if [ "$1" = "error" ]; then echo "$2" >&2; else echo "$2"; fi; }
+  export -f log
+
+  export K8S_NAMESPACE="test-namespace"
+  export SCOPE_ID="scope-123"
+  export KILL_STATE="$BATS_TEST_TMPDIR/deleted"
+  rm -f "$KILL_STATE"
+
+  export CONTEXT='{
+    "parameters": {
+      "deployment_id": "deploy-456",
+      "instance_name": "my-pod-abc123"
+    },
+    "tags": {
+      "scope_id": "scope-123"
+    },
+    "providers": {
+      "container-orchestration": {
+        "cluster": {
+          "namespace": "test-namespace"
+        }
+      }
+    }
+  }'
+
+  # Default mock: pod owned by a Job that belongs to the expected CronJob.
+  # A state file flips the plain (no -o) pod existence check to "gone" once
+  # `kubectl delete` has run, exercising the happy termination path.
+  kubectl() {
+    case "$1" in
+      get)
+        case "$2" in
+          pod)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"phase"* ]]; then
+                echo "Running"
+              elif [[ "$*" == *"nodeName"* ]]; then
+                echo "node-1"
+              elif [[ "$*" == *"startTime"* ]]; then
+                echo "2024-01-01T00:00:00Z"
+              elif [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456-abc"
+              fi
+              return 0
+            fi
+            # existence check (no -o): gone once deleted
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+          job)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456"
+              elif [[ "$*" == *"active"* ]]; then
+                echo "0"
+              elif [[ "$*" == *"succeeded"* ]]; then
+                echo "1"
+              elif [[ "$*" == *"failed"* ]]; then
+                echo "0"
+              fi
+              return 0
+            fi
+            return 0
+            ;;
+        esac
+        ;;
+      delete)
+        touch "$KILL_STATE"
+        return 0
+        ;;
+      wait)
+        return 0
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+}
+
+teardown() {
+  unset CONTEXT
+  unset -f kubectl
+}
+
+# =============================================================================
+# Success Case
+# =============================================================================
+@test "kill_instances: successfully kills job pod with correct logging" {
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 0 ]
+  # Start message
+  assert_contains "$output" "🔍 Starting instance kill operation..."
+  # Parameter display
+  assert_contains "$output" "📋 Deployment ID: deploy-456"
+  assert_contains "$output" "📋 Instance name: my-pod-abc123"
+  assert_contains "$output" "📋 Scope ID: scope-123"
+  assert_contains "$output" "📋 Namespace: test-namespace"
+  # Pod verification
+  assert_contains "$output" "🔍 Verifying pod exists..."
+  assert_contains "$output" "📋 Fetching pod details..."
+  assert_contains "$output" "📋 Pod: my-pod-abc123 | Status: Running | Node: node-1 | Started: 2024-01-01T00:00:00Z"
+  # Ownership (Job -> CronJob)
+  assert_contains "$output" "📋 Pod ownership: Job=job-scope-123-deploy-456-abc -> CronJob=job-scope-123-deploy-456"
+  # Delete operation
+  assert_contains "$output" "📝 Deleting pod my-pod-abc123 with 30s grace period..."
+  assert_contains "$output" "📝 Waiting for pod termination..."
+  assert_contains "$output" "✅ Pod successfully terminated and removed"
+  # Job status
+  assert_contains "$output" "📋 Checking job status after pod deletion..."
+  assert_contains "$output" "📋 Job job-scope-123-deploy-456-abc: active=0, succeeded=1, failed=0"
+  # Completion
+  assert_contains "$output" "✨ Instance kill operation completed for my-pod-abc123"
+}
+
+# =============================================================================
+# Error Cases
+# =============================================================================
+@test "kill_instances: fails with troubleshooting when deployment_id missing" {
+  export CONTEXT='{
+    "parameters": {
+      "instance_name": "my-pod-abc123"
+    }
+  }'
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ deployment_id parameter not found"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "Parameter not provided in action request"
+  assert_contains "$output" "🔧 How to fix:"
+  assert_contains "$output" "Ensure deployment_id is passed in the action parameters"
+}
+
+@test "kill_instances: fails with troubleshooting when instance_name missing" {
+  export CONTEXT='{
+    "parameters": {
+      "deployment_id": "deploy-456"
+    }
+  }'
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ instance_name parameter not found"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "Parameter not provided in action request"
+  assert_contains "$output" "🔧 How to fix:"
+  assert_contains "$output" "Ensure instance_name is passed in the action parameters"
+}
+
+@test "kill_instances: fails with troubleshooting when scope_id missing" {
+  export CONTEXT='{
+    "parameters": {
+      "deployment_id": "deploy-456",
+      "instance_name": "my-pod-abc123"
+    }
+  }'
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ scope_id not found in context"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "Context missing scope information"
+  assert_contains "$output" "🔧 How to fix:"
+  assert_contains "$output" "Verify the action is invoked with proper scope context"
+}
+
+@test "kill_instances: fails with troubleshooting when pod not found" {
+  kubectl() {
+    case "$1" in
+      get)
+        if [[ "$2" == "pod" ]] && [[ "$*" != *"-o"* ]]; then
+          return 1
+        fi
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 1 ]
+  assert_contains "$output" "❌ Pod my-pod-abc123 not found in namespace test-namespace"
+  assert_contains "$output" "💡 Possible causes:"
+  assert_contains "$output" "Pod was already terminated"
+  assert_contains "$output" "🔧 How to fix:"
+  assert_contains "$output" "kubectl get pods"
+}
+
+# =============================================================================
+# Warning Cases
+# =============================================================================
+@test "kill_instances: warns when pod belongs to a different scheduled task" {
+  kubectl() {
+    case "$1" in
+      get)
+        case "$2" in
+          pod)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"phase"* ]]; then
+                echo "Running"
+              elif [[ "$*" == *"nodeName"* ]]; then
+                echo "node-1"
+              elif [[ "$*" == *"startTime"* ]]; then
+                echo "2024-01-01T00:00:00Z"
+              elif [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456-abc"
+              fi
+              return 0
+            fi
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+          job)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-different-deploy"  # Different CronJob
+              elif [[ "$*" == *"active"* ]]; then
+                echo "0"
+              fi
+              return 0
+            fi
+            return 0
+            ;;
+        esac
+        ;;
+      delete)
+        touch "$KILL_STATE"
+        return 0
+        ;;
+      wait)
+        return 0
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "⚠️  Pod does not belong to expected scheduled task job-scope-123-deploy-456 (continuing anyway)"
+}
+
+@test "kill_instances: warns when pod ownership cannot be verified" {
+  kubectl() {
+    case "$1" in
+      get)
+        case "$2" in
+          pod)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"phase"* ]]; then
+                echo "Running"
+              elif [[ "$*" == *"nodeName"* ]]; then
+                echo "node-1"
+              elif [[ "$*" == *"startTime"* ]]; then
+                echo "2024-01-01T00:00:00Z"
+              elif [[ "$*" == *"ownerReferences"* ]]; then
+                echo ""  # Bare pod, no owner
+              fi
+              return 0
+            fi
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+          job)
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+        esac
+        ;;
+      delete)
+        touch "$KILL_STATE"
+        return 0
+        ;;
+      wait)
+        return 0
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "⚠️  Could not verify pod ownership"
+  # With no owner Job, the post-deletion job lookup cannot resolve one either
+  assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
+}
+
+@test "kill_instances: warns when pod still exists after deletion" {
+  kubectl() {
+    case "$1" in
+      get)
+        case "$2" in
+          pod)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"phase"* ]]; then
+                echo "Terminating"
+              elif [[ "$*" == *"nodeName"* ]]; then
+                echo "node-1"
+              elif [[ "$*" == *"startTime"* ]]; then
+                echo "2024-01-01T00:00:00Z"
+              elif [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456-abc"
+              fi
+              return 0
+            fi
+            return 0  # Pod still exists
+            ;;
+          job)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456"
+              elif [[ "$*" == *"active"* ]]; then
+                echo "1"
+              fi
+              return 0
+            fi
+            return 0
+            ;;
+        esac
+        ;;
+      delete)
+        return 0
+        ;;
+      wait)
+        return 1  # Timeout
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "⚠️  Pod deletion timeout reached"
+  assert_contains "$output" "⚠️  Pod still exists after deletion attempt (status: Terminating)"
+  # Job still active -> replacement pod may be started
+  assert_contains "$output" "📋 Checking job status after pod deletion..."
+  assert_contains "$output" "📋 Job job-scope-123-deploy-456-abc: active=1, succeeded=0, failed=0"
+  assert_contains "$output" "📋 Job is still active; Kubernetes may start a replacement pod (backoffLimit permitting)"
+}
+
+@test "kill_instances: warns when job already completed after pod deletion" {
+  kubectl() {
+    case "$1" in
+      get)
+        case "$2" in
+          pod)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"phase"* ]]; then
+                echo "Running"
+              elif [[ "$*" == *"nodeName"* ]]; then
+                echo "node-1"
+              elif [[ "$*" == *"startTime"* ]]; then
+                echo "2024-01-01T00:00:00Z"
+              elif [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456-abc"
+              fi
+              return 0
+            fi
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+          job)
+            if [[ "$*" == *"-o jsonpath"* ]]; then
+              if [[ "$*" == *"ownerReferences"* ]]; then
+                echo "job-scope-123-deploy-456"
+              fi
+              return 0
+            fi
+            # Job existence check: gone once the pod was deleted
+            [[ -f "$KILL_STATE" ]] && return 1
+            return 0
+            ;;
+        esac
+        ;;
+      delete)
+        touch "$KILL_STATE"
+        return 0
+        ;;
+      wait)
+        return 0
+        ;;
+    esac
+    return 0
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
+}
diff --git a/scheduled_task/deployment/tests/workflow_overrides.bats b/scheduled_task/deployment/tests/workflow_overrides.bats
index 591d00db..4bbe17c1 100644
--- a/scheduled_task/deployment/tests/workflow_overrides.bats
+++ b/scheduled_task/deployment/tests/workflow_overrides.bats
@@ -34,3 +34,43 @@ setup() {
   assert_equal "$status" "0"
   assert_contains "$output" "action: skip"
 }
+
+# =============================================================================
+# kill instances
+#
+# scheduled_task job pods are owned by Job -> CronJob, so the base k8s kill
+# script (which reasons about Deployment/ReplicaSet ownership) is replaced by a
+# job-aware script. The base must keep the step name so the overlay keeps
+# matching it after upstream changes.
+# =============================================================================
+@test "base k8s deployment kill_instances workflow declares 'kill instances' step" {
+  run grep -A 2 "name: kill instances" "$PROJECT_ROOT/k8s/deployment/workflows/kill_instances.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "type: script"
+  assert_contains "$output" "deployment/kill_instances"
+}
+
+@test "scheduled_task deployment kill_instances overlay replaces 'kill instances' with its own script" {
+  run grep -A 3 "name: kill instances" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_instances.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "action: replace"
+  assert_contains "$output" "\$OVERRIDES_PATH/deployment/kill_instances"
+}
+
+@test "base k8s deployment kill_instances workflow declares 'load logging' step" {
+  run grep -A 2 "name: load logging" "$PROJECT_ROOT/k8s/deployment/workflows/kill_instances.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "type: script"
+  assert_contains "$output" "logging"
+}
+
+@test "scheduled_task deployment kill_instances overlay loads its own logging function" {
+  run grep -A 3 "name: load logging" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_instances.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "action: replace"
+  assert_contains "$output" "\$OVERRIDES_PATH/logging"
+}
diff --git a/scheduled_task/deployment/workflows/kill_instances.yaml b/scheduled_task/deployment/workflows/kill_instances.yaml
new file mode 100644
index 00000000..c3b17eb0
--- /dev/null
+++ b/scheduled_task/deployment/workflows/kill_instances.yaml
@@ -0,0 +1,17 @@
+include:
+  - "$SERVICE_PATH/values.yaml"
+steps:
+  - name: load logging
+    type: script
+    action: replace
+    file: "$OVERRIDES_PATH/logging"
+    output:
+      - name: log
+        type: function
+        parameters:
+          level: string
+          message: string
+  - name: kill instances
+    type: script
+    action: replace
+    file: "$OVERRIDES_PATH/deployment/kill_instances"
diff --git a/scheduled_task/specs/actions/kill-instances.json.tpl b/scheduled_task/specs/actions/kill-instances.json.tpl
new file mode 100644
index 00000000..c1351f94
--- /dev/null
+++ b/scheduled_task/specs/actions/kill-instances.json.tpl
@@ -0,0 +1,18 @@
+{
+  "name": "Kill instances",
+  "type": "custom",
+  "icon": "material-symbols:delete-outline",
+  "results": {
+    "schema": { "type": "object", "required": [], "properties": {} },
+    "values": {}
+  },
+  "service_specification_id": "{{ env.Getenv "SERVICE_SPECIFICATION_ID" }}",
+  "parameters": {
+    "schema": { "type": "object", "required": [], "properties": {} },
+    "values": {}
+  },
+  "annotations": {
+    "show_on": ["performance"],
+    "runs_over": "scope"
+  }
+}
\ No newline at end of file

From 348d047ba8b0769ba3e77e11018d5336f65f82e9 Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 14:22:34 -0300
Subject: [PATCH 4/7] refactor: rename scheduled task kill action to
 kill-job-execution

---
 CHANGELOG.md                                  |  2 +-
 .../{kill_instances => kill_job_execution}    |  0
 ...instances.bats => kill_job_execution.bats} | 38 +++++++++---------
 .../deployment/tests/workflow_overrides.bats  | 40 +++++++++----------
 ...instances.yaml => kill_job_execution.yaml} |  6 +--
 ...s.json.tpl => kill-job-execution.json.tpl} |  2 +-
 service/deployment/entrypoint                 | 20 +++++++++-
 service/deployment/tests/entrypoint.bats      | 26 ++++++++++++
 8 files changed, 86 insertions(+), 48 deletions(-)
 rename scheduled_task/deployment/{kill_instances => kill_job_execution} (100%)
 rename scheduled_task/deployment/tests/{kill_instances.bats => kill_job_execution.bats} (90%)
 rename scheduled_task/deployment/workflows/{kill_instances.yaml => kill_job_execution.yaml} (68%)
 rename scheduled_task/specs/actions/{kill-instances.json.tpl => kill-job-execution.json.tpl} (93%)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 9a7b51bd..1e6ca99b 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,7 +7,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 - 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 instances" action to scheduled task scopes to terminate an individual running job instance
+- Add "Kill job execution" action to scheduled task scopes to terminate an individual running job execution
 
 ## [1.13.0] - 2026-07-10
 - Add support to get AWS credentials via assume role
diff --git a/scheduled_task/deployment/kill_instances b/scheduled_task/deployment/kill_job_execution
similarity index 100%
rename from scheduled_task/deployment/kill_instances
rename to scheduled_task/deployment/kill_job_execution
diff --git a/scheduled_task/deployment/tests/kill_instances.bats b/scheduled_task/deployment/tests/kill_job_execution.bats
similarity index 90%
rename from scheduled_task/deployment/tests/kill_instances.bats
rename to scheduled_task/deployment/tests/kill_job_execution.bats
index 6f1d86e4..94e1569b 100644
--- a/scheduled_task/deployment/tests/kill_instances.bats
+++ b/scheduled_task/deployment/tests/kill_job_execution.bats
@@ -1,6 +1,6 @@
 #!/usr/bin/env bats
 # =============================================================================
-# Unit tests for scheduled_task deployment/kill_instances - job pod termination
+# Unit tests for scheduled_task deployment/kill_job_execution - job pod termination
 #
 # scheduled_task pods are owned by Job -> CronJob (job-<scope>-<deployment>),
 # unlike the base k8s scope where pods are owned by ReplicaSet -> Deployment.
@@ -96,8 +96,8 @@ teardown() {
 # =============================================================================
 # Success Case
 # =============================================================================
-@test "kill_instances: successfully kills job pod with correct logging" {
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+@test "kill_job_execution: successfully kills job pod with correct logging" {
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
   # Start message
@@ -127,14 +127,14 @@ teardown() {
 # =============================================================================
 # Error Cases
 # =============================================================================
-@test "kill_instances: fails with troubleshooting when deployment_id missing" {
+@test "kill_job_execution: fails with troubleshooting when deployment_id missing" {
   export CONTEXT='{
     "parameters": {
       "instance_name": "my-pod-abc123"
     }
   }'
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 1 ]
   assert_contains "$output" "❌ deployment_id parameter not found"
@@ -144,14 +144,14 @@ teardown() {
   assert_contains "$output" "Ensure deployment_id is passed in the action parameters"
 }
 
-@test "kill_instances: fails with troubleshooting when instance_name missing" {
+@test "kill_job_execution: fails with troubleshooting when instance_name missing" {
   export CONTEXT='{
     "parameters": {
       "deployment_id": "deploy-456"
     }
   }'
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 1 ]
   assert_contains "$output" "❌ instance_name parameter not found"
@@ -161,7 +161,7 @@ teardown() {
   assert_contains "$output" "Ensure instance_name is passed in the action parameters"
 }
 
-@test "kill_instances: fails with troubleshooting when scope_id missing" {
+@test "kill_job_execution: fails with troubleshooting when scope_id missing" {
   export CONTEXT='{
     "parameters": {
       "deployment_id": "deploy-456",
@@ -169,7 +169,7 @@ teardown() {
     }
   }'
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 1 ]
   assert_contains "$output" "❌ scope_id not found in context"
@@ -179,7 +179,7 @@ teardown() {
   assert_contains "$output" "Verify the action is invoked with proper scope context"
 }
 
-@test "kill_instances: fails with troubleshooting when pod not found" {
+@test "kill_job_execution: fails with troubleshooting when pod not found" {
   kubectl() {
     case "$1" in
       get)
@@ -192,7 +192,7 @@ teardown() {
   }
   export -f kubectl
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 1 ]
   assert_contains "$output" "❌ Pod my-pod-abc123 not found in namespace test-namespace"
@@ -205,7 +205,7 @@ teardown() {
 # =============================================================================
 # Warning Cases
 # =============================================================================
-@test "kill_instances: warns when pod belongs to a different scheduled task" {
+@test "kill_job_execution: warns when pod belongs to a different scheduled task" {
   kubectl() {
     case "$1" in
       get)
@@ -251,13 +251,13 @@ teardown() {
   }
   export -f kubectl
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
   assert_contains "$output" "⚠️  Pod does not belong to expected scheduled task job-scope-123-deploy-456 (continuing anyway)"
 }
 
-@test "kill_instances: warns when pod ownership cannot be verified" {
+@test "kill_job_execution: warns when pod ownership cannot be verified" {
   kubectl() {
     case "$1" in
       get)
@@ -296,7 +296,7 @@ teardown() {
   }
   export -f kubectl
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
   assert_contains "$output" "⚠️  Could not verify pod ownership"
@@ -304,7 +304,7 @@ teardown() {
   assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
 }
 
-@test "kill_instances: warns when pod still exists after deletion" {
+@test "kill_job_execution: warns when pod still exists after deletion" {
   kubectl() {
     case "$1" in
       get)
@@ -348,7 +348,7 @@ teardown() {
   }
   export -f kubectl
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
   assert_contains "$output" "⚠️  Pod deletion timeout reached"
@@ -359,7 +359,7 @@ teardown() {
   assert_contains "$output" "📋 Job is still active; Kubernetes may start a replacement pod (backoffLimit permitting)"
 }
 
-@test "kill_instances: warns when job already completed after pod deletion" {
+@test "kill_job_execution: warns when job already completed after pod deletion" {
   kubectl() {
     case "$1" in
       get)
@@ -405,7 +405,7 @@ teardown() {
   }
   export -f kubectl
 
-  run bash "$BATS_TEST_DIRNAME/../kill_instances"
+  run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
   assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
diff --git a/scheduled_task/deployment/tests/workflow_overrides.bats b/scheduled_task/deployment/tests/workflow_overrides.bats
index 4bbe17c1..69aa520f 100644
--- a/scheduled_task/deployment/tests/workflow_overrides.bats
+++ b/scheduled_task/deployment/tests/workflow_overrides.bats
@@ -36,41 +36,37 @@ setup() {
 }
 
 # =============================================================================
-# kill instances
+# kill job execution
 #
-# scheduled_task job pods are owned by Job -> CronJob, so the base k8s kill
-# script (which reasons about Deployment/ReplicaSet ownership) is replaced by a
-# job-aware script. The base must keep the step name so the overlay keeps
-# matching it after upstream changes.
+# scheduled_task pods are owned by Job -> CronJob, so this action ships a
+# STANDALONE deployment workflow (not an override of a k8s base) with its own
+# steps. The deployment entrypoint falls back to the override-provided workflow
+# because k8s has no kill_job_execution counterpart.
 # =============================================================================
-@test "base k8s deployment kill_instances workflow declares 'kill instances' step" {
-  run grep -A 2 "name: kill instances" "$PROJECT_ROOT/k8s/deployment/workflows/kill_instances.yaml"
+@test "kill_job_execution has no base k8s workflow (it is standalone)" {
+  run test -f "$PROJECT_ROOT/k8s/deployment/workflows/kill_job_execution.yaml"
 
-  assert_equal "$status" "0"
-  assert_contains "$output" "type: script"
-  assert_contains "$output" "deployment/kill_instances"
+  assert_equal "$status" "1"
 }
 
-@test "scheduled_task deployment kill_instances overlay replaces 'kill instances' with its own script" {
-  run grep -A 3 "name: kill instances" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_instances.yaml"
+@test "scheduled_task deployment kill_job_execution workflow loads its own logging function" {
+  run grep -A 3 "name: load logging" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_job_execution.yaml"
 
   assert_equal "$status" "0"
-  assert_contains "$output" "action: replace"
-  assert_contains "$output" "\$OVERRIDES_PATH/deployment/kill_instances"
+  assert_contains "$output" "type: script"
+  assert_contains "$output" "\$OVERRIDES_PATH/logging"
 }
 
-@test "base k8s deployment kill_instances workflow declares 'load logging' step" {
-  run grep -A 2 "name: load logging" "$PROJECT_ROOT/k8s/deployment/workflows/kill_instances.yaml"
+@test "scheduled_task deployment kill_job_execution workflow runs its own kill script" {
+  run grep -A 2 "name: kill job execution" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_job_execution.yaml"
 
   assert_equal "$status" "0"
   assert_contains "$output" "type: script"
-  assert_contains "$output" "logging"
+  assert_contains "$output" "\$OVERRIDES_PATH/deployment/kill_job_execution"
 }
 
-@test "scheduled_task deployment kill_instances overlay loads its own logging function" {
-  run grep -A 3 "name: load logging" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_instances.yaml"
+@test "scheduled_task deployment kill_job_execution workflow is not an override (no action: replace)" {
+  run grep -c "action: replace" "$PROJECT_ROOT/scheduled_task/deployment/workflows/kill_job_execution.yaml"
 
-  assert_equal "$status" "0"
-  assert_contains "$output" "action: replace"
-  assert_contains "$output" "\$OVERRIDES_PATH/logging"
+  assert_contains "$output" "0"
 }
diff --git a/scheduled_task/deployment/workflows/kill_instances.yaml b/scheduled_task/deployment/workflows/kill_job_execution.yaml
similarity index 68%
rename from scheduled_task/deployment/workflows/kill_instances.yaml
rename to scheduled_task/deployment/workflows/kill_job_execution.yaml
index c3b17eb0..ef8de4bd 100644
--- a/scheduled_task/deployment/workflows/kill_instances.yaml
+++ b/scheduled_task/deployment/workflows/kill_job_execution.yaml
@@ -3,7 +3,6 @@ include:
 steps:
   - name: load logging
     type: script
-    action: replace
     file: "$OVERRIDES_PATH/logging"
     output:
       - name: log
@@ -11,7 +10,6 @@ steps:
         parameters:
           level: string
           message: string
-  - name: kill instances
+  - name: kill job execution
     type: script
-    action: replace
-    file: "$OVERRIDES_PATH/deployment/kill_instances"
+    file: "$OVERRIDES_PATH/deployment/kill_job_execution"
diff --git a/scheduled_task/specs/actions/kill-instances.json.tpl b/scheduled_task/specs/actions/kill-job-execution.json.tpl
similarity index 93%
rename from scheduled_task/specs/actions/kill-instances.json.tpl
rename to scheduled_task/specs/actions/kill-job-execution.json.tpl
index c1351f94..405a6f85 100644
--- a/scheduled_task/specs/actions/kill-instances.json.tpl
+++ b/scheduled_task/specs/actions/kill-job-execution.json.tpl
@@ -1,5 +1,5 @@
 {
-  "name": "Kill instances",
+  "name": "Kill job execution",
   "type": "custom",
   "icon": "material-symbols:delete-outline",
   "results": {
diff --git a/service/deployment/entrypoint b/service/deployment/entrypoint
index 9a1ee9b1..58339602 100755
--- a/service/deployment/entrypoint
+++ b/service/deployment/entrypoint
@@ -29,6 +29,9 @@ case "$SERVICE_ACTION" in
   "kill-instances")
     ACTION_TO_EXECUTE="kill_instances"
     ;;
+  "kill-job-execution")
+    ACTION_TO_EXECUTE="kill_job_execution"
+    ;;
   *)
     echo "Unknown action: $SERVICE_ACTION"
     exit 1
@@ -37,9 +40,24 @@ esac
 
 WORKFLOW_PATH="$SERVICE_PATH/deployment/workflows/$ACTION_TO_EXECUTE.yaml"
 
+# Fall back to an override-provided workflow when the base scope does not define
+# one. This lets an override scope (e.g. scheduled_task) ship a standalone
+# deployment workflow that has no base counterpart.
+if [[ ! -f "$WORKFLOW_PATH" ]]; then
+  IFS=',' read -ra OVERRIDE_PATHS <<< "$OVERRIDES_PATH"
+  for path in "${OVERRIDE_PATHS[@]}"; do
+    path=$(echo "$path" | xargs)
+    CANDIDATE_WORKFLOW_PATH="$path/deployment/workflows/$ACTION_TO_EXECUTE.yaml"
+    if [[ -f "$CANDIDATE_WORKFLOW_PATH" ]]; then
+      WORKFLOW_PATH="$CANDIDATE_WORKFLOW_PATH"
+      break
+    fi
+  done
+fi
+
 NEEDS_PARAMS=true
 case "$SERVICE_ACTION" in
-  "switch-traffic"|"kill-instances"|"diagnose-deployment")
+  "switch-traffic"|"kill-instances"|"kill-job-execution"|"diagnose-deployment")
     NEEDS_PARAMS=false
     ;;
 esac
diff --git a/service/deployment/tests/entrypoint.bats b/service/deployment/tests/entrypoint.bats
index fac10a8f..1ea13eac 100644
--- a/service/deployment/tests/entrypoint.bats
+++ b/service/deployment/tests/entrypoint.bats
@@ -79,6 +79,32 @@ mock_np() {
   assert_contains "$output" "--no-params"
 }
 
+@test "deployment entrypoint: kill-job-execution includes --no-params when CLI supports it" {
+  mock_np
+  export SERVICE_ACTION="kill-job-execution"
+  mkdir -p "$BATS_TEST_TMPDIR/override/deployment/workflows"
+  touch "$BATS_TEST_TMPDIR/override/deployment/workflows/kill_job_execution.yaml"
+  export OVERRIDES_PATH="$BATS_TEST_TMPDIR/override"
+
+  run bash "$BATS_TEST_DIRNAME/../entrypoint"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "--no-params"
+}
+
+@test "deployment entrypoint: kill-job-execution falls back to override workflow when base is missing" {
+  mock_np
+  export SERVICE_ACTION="kill-job-execution"
+  mkdir -p "$BATS_TEST_TMPDIR/override/deployment/workflows"
+  touch "$BATS_TEST_TMPDIR/override/deployment/workflows/kill_job_execution.yaml"
+  export OVERRIDES_PATH="$BATS_TEST_TMPDIR/override"
+
+  run bash "$BATS_TEST_DIRNAME/../entrypoint"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "--workflow $BATS_TEST_TMPDIR/override/deployment/workflows/kill_job_execution.yaml"
+}
+
 # =============================================================================
 # Actions that SHOULD NOT include --no-params
 # =============================================================================

From 358b73ea43d0046302fa82b3418d236cd95227b0 Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 14:39:18 -0300
Subject: [PATCH 5/7] refactor: reword kill-job-execution logs to job execution
 terminology

---
 scheduled_task/deployment/kill_job_execution  | 44 ++++++++---------
 .../deployment/tests/kill_job_execution.bats  | 48 +++++++++----------
 2 files changed, 46 insertions(+), 46 deletions(-)

diff --git a/scheduled_task/deployment/kill_job_execution b/scheduled_task/deployment/kill_job_execution
index 4a0bb22d..6314c168 100755
--- a/scheduled_task/deployment/kill_job_execution
+++ b/scheduled_task/deployment/kill_job_execution
@@ -3,7 +3,7 @@
 set -euo pipefail
 
 
-log debug "🔍 Starting instance kill operation..."
+log debug "🔍 Starting job execution kill operation..."
 
 DEPLOYMENT_ID=$(echo "$CONTEXT" | jq -r '.parameters.deployment_id // .notification.parameters.deployment_id // empty')
 INSTANCE_NAME=$(echo "$CONTEXT" | jq -r '.parameters.instance_name // .notification.parameters.instance_name // empty')
@@ -37,7 +37,7 @@ if [[ -z "$INSTANCE_NAME" ]]; then
 fi
 
 log debug "📋 Deployment ID: $DEPLOYMENT_ID"
-log debug "📋 Instance name: $INSTANCE_NAME"
+log debug "📋 Job execution: $INSTANCE_NAME"
 
 SCOPE_ID=$(echo "$CONTEXT" | jq -r '.tags.scope_id // .scope.id // .notification.tags.scope_id // empty')
 
@@ -62,53 +62,53 @@ fi
 log debug "📋 Scope ID: $SCOPE_ID"
 log debug "📋 Namespace: $K8S_NAMESPACE"
 
-log debug "🔍 Verifying pod exists..."
+log debug "🔍 Verifying job execution exists..."
 if ! kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
-    log error "❌ Pod $INSTANCE_NAME not found in namespace $K8S_NAMESPACE"
+    log error "❌ Job execution $INSTANCE_NAME not found in namespace $K8S_NAMESPACE"
     log error "💡 Possible causes:"
-    log error "   - Pod was already terminated"
-    log error "   - Pod name is incorrect"
-    log error "   - Pod exists in a different namespace"
+    log error "   - The job execution was already terminated"
+    log error "   - The job execution name is incorrect"
+    log error "   - The job execution runs in a different namespace"
     log error "🔧 How to fix:"
-    log error "   - List pods: kubectl get pods -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID"
+    log error "   - List job executions: kubectl get pods -n $K8S_NAMESPACE -l scope_id=$SCOPE_ID"
     exit 1
 fi
 
-log debug "📋 Fetching pod details..."
+log debug "📋 Fetching job execution details..."
 POD_STATUS=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.phase}')
 POD_NODE=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.spec.nodeName}')
 POD_START_TIME=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.startTime}')
 
-log debug "📋 Pod: $INSTANCE_NAME | Status: $POD_STATUS | Node: $POD_NODE | Started: $POD_START_TIME"
+log debug "📋 Status: $POD_STATUS | Node: $POD_NODE | Started: $POD_START_TIME"
 
 CRONJOB_NAME="job-$SCOPE_ID-$DEPLOYMENT_ID"
 
 POD_JOB=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.metadata.ownerReferences[0].name}' 2>/dev/null || echo "")
 if [[ -n "$POD_JOB" ]]; then
     JOB_CRONJOB=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.metadata.ownerReferences[0].name}' 2>/dev/null || echo "")
-    log debug "📋 Pod ownership: Job=$POD_JOB -> CronJob=$JOB_CRONJOB"
+    log debug "📋 Ownership: Job=$POD_JOB -> CronJob=$JOB_CRONJOB"
 
     if [[ "$JOB_CRONJOB" != "$CRONJOB_NAME" ]]; then
-        log warn "⚠️  Pod does not belong to expected scheduled task $CRONJOB_NAME (continuing anyway)"
+        log warn "⚠️  Job execution does not belong to expected scheduled task $CRONJOB_NAME (continuing anyway)"
     fi
 else
-    log warn "⚠️  Could not verify pod ownership"
+    log warn "⚠️  Could not verify job execution ownership"
 fi
 
-log debug "📝 Deleting pod $INSTANCE_NAME with 30s grace period..."
+log debug "📝 Deleting job execution $INSTANCE_NAME with 30s grace period..."
 kubectl delete pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" --grace-period=30
 
-log debug "📝 Waiting for pod termination..."
-kubectl wait --for=delete pod/"$INSTANCE_NAME" -n "$K8S_NAMESPACE" --timeout=60s || log warn "⚠️  Pod deletion timeout reached"
+log debug "📝 Waiting for job execution termination..."
+kubectl wait --for=delete pod/"$INSTANCE_NAME" -n "$K8S_NAMESPACE" --timeout=60s || log warn "⚠️  Job execution deletion timeout reached"
 
 if kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
     POD_STATUS_AFTER=$(kubectl get pod "$INSTANCE_NAME" -n "$K8S_NAMESPACE" -o jsonpath='{.status.phase}')
-    log warn "⚠️  Pod still exists after deletion attempt (status: $POD_STATUS_AFTER)"
+    log warn "⚠️  Job execution still exists after deletion attempt (status: $POD_STATUS_AFTER)"
 else
-    log info "✅ Pod successfully terminated and removed"
+    log info "✅ Job execution successfully terminated and removed"
 fi
 
-log debug "📋 Checking job status after pod deletion..."
+log debug "📋 Checking parent job status after execution deletion..."
 if [[ -n "$POD_JOB" ]] && kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" >/dev/null 2>&1; then
     JOB_ACTIVE=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.status.active}')
     JOB_SUCCEEDED=$(kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" -o jsonpath='{.status.succeeded}')
@@ -117,10 +117,10 @@ if [[ -n "$POD_JOB" ]] && kubectl get job "$POD_JOB" -n "$K8S_NAMESPACE" >/dev/n
     log debug "📋 Job $POD_JOB: active=${JOB_ACTIVE:-0}, succeeded=${JOB_SUCCEEDED:-0}, failed=${JOB_FAILED:-0}"
 
     if [[ "${JOB_ACTIVE:-0}" -gt 0 ]]; then
-        log debug "📋 Job is still active; Kubernetes may start a replacement pod (backoffLimit permitting)"
+        log debug "📋 Job is still active; Kubernetes may start a replacement execution (backoffLimit permitting)"
     fi
 else
-    log warn "⚠️  Job for pod not found (it may have already completed)"
+    log warn "⚠️  Parent job not found (it may have already completed)"
 fi
 
-log info "✨ Instance kill operation completed for $INSTANCE_NAME"
+log info "✨ Job execution kill operation completed for $INSTANCE_NAME"
diff --git a/scheduled_task/deployment/tests/kill_job_execution.bats b/scheduled_task/deployment/tests/kill_job_execution.bats
index 94e1569b..e0a44034 100644
--- a/scheduled_task/deployment/tests/kill_job_execution.bats
+++ b/scheduled_task/deployment/tests/kill_job_execution.bats
@@ -101,27 +101,27 @@ teardown() {
 
   [ "$status" -eq 0 ]
   # Start message
-  assert_contains "$output" "🔍 Starting instance kill operation..."
+  assert_contains "$output" "🔍 Starting job execution kill operation..."
   # Parameter display
   assert_contains "$output" "📋 Deployment ID: deploy-456"
-  assert_contains "$output" "📋 Instance name: my-pod-abc123"
+  assert_contains "$output" "📋 Job execution: my-pod-abc123"
   assert_contains "$output" "📋 Scope ID: scope-123"
   assert_contains "$output" "📋 Namespace: test-namespace"
-  # Pod verification
-  assert_contains "$output" "🔍 Verifying pod exists..."
-  assert_contains "$output" "📋 Fetching pod details..."
-  assert_contains "$output" "📋 Pod: my-pod-abc123 | Status: Running | Node: node-1 | Started: 2024-01-01T00:00:00Z"
+  # Job execution verification
+  assert_contains "$output" "🔍 Verifying job execution exists..."
+  assert_contains "$output" "📋 Fetching job execution details..."
+  assert_contains "$output" "📋 Status: Running | Node: node-1 | Started: 2024-01-01T00:00:00Z"
   # Ownership (Job -> CronJob)
-  assert_contains "$output" "📋 Pod ownership: Job=job-scope-123-deploy-456-abc -> CronJob=job-scope-123-deploy-456"
+  assert_contains "$output" "📋 Ownership: Job=job-scope-123-deploy-456-abc -> CronJob=job-scope-123-deploy-456"
   # Delete operation
-  assert_contains "$output" "📝 Deleting pod my-pod-abc123 with 30s grace period..."
-  assert_contains "$output" "📝 Waiting for pod termination..."
-  assert_contains "$output" "✅ Pod successfully terminated and removed"
-  # Job status
-  assert_contains "$output" "📋 Checking job status after pod deletion..."
+  assert_contains "$output" "📝 Deleting job execution my-pod-abc123 with 30s grace period..."
+  assert_contains "$output" "📝 Waiting for job execution termination..."
+  assert_contains "$output" "✅ Job execution successfully terminated and removed"
+  # Parent job status
+  assert_contains "$output" "📋 Checking parent job status after execution deletion..."
   assert_contains "$output" "📋 Job job-scope-123-deploy-456-abc: active=0, succeeded=1, failed=0"
   # Completion
-  assert_contains "$output" "✨ Instance kill operation completed for my-pod-abc123"
+  assert_contains "$output" "✨ Job execution kill operation completed for my-pod-abc123"
 }
 
 # =============================================================================
@@ -195,9 +195,9 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 1 ]
-  assert_contains "$output" "❌ Pod my-pod-abc123 not found in namespace test-namespace"
+  assert_contains "$output" "❌ Job execution my-pod-abc123 not found in namespace test-namespace"
   assert_contains "$output" "💡 Possible causes:"
-  assert_contains "$output" "Pod was already terminated"
+  assert_contains "$output" "The job execution was already terminated"
   assert_contains "$output" "🔧 How to fix:"
   assert_contains "$output" "kubectl get pods"
 }
@@ -254,7 +254,7 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "⚠️  Pod does not belong to expected scheduled task job-scope-123-deploy-456 (continuing anyway)"
+  assert_contains "$output" "⚠️  Job execution does not belong to expected scheduled task job-scope-123-deploy-456 (continuing anyway)"
 }
 
 @test "kill_job_execution: warns when pod ownership cannot be verified" {
@@ -299,9 +299,9 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "⚠️  Could not verify pod ownership"
+  assert_contains "$output" "⚠️  Could not verify job execution ownership"
   # With no owner Job, the post-deletion job lookup cannot resolve one either
-  assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
+  assert_contains "$output" "⚠️  Parent job not found (it may have already completed)"
 }
 
 @test "kill_job_execution: warns when pod still exists after deletion" {
@@ -351,12 +351,12 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "⚠️  Pod deletion timeout reached"
-  assert_contains "$output" "⚠️  Pod still exists after deletion attempt (status: Terminating)"
-  # Job still active -> replacement pod may be started
-  assert_contains "$output" "📋 Checking job status after pod deletion..."
+  assert_contains "$output" "⚠️  Job execution deletion timeout reached"
+  assert_contains "$output" "⚠️  Job execution still exists after deletion attempt (status: Terminating)"
+  # Job still active -> replacement execution may be started
+  assert_contains "$output" "📋 Checking parent job status after execution deletion..."
   assert_contains "$output" "📋 Job job-scope-123-deploy-456-abc: active=1, succeeded=0, failed=0"
-  assert_contains "$output" "📋 Job is still active; Kubernetes may start a replacement pod (backoffLimit permitting)"
+  assert_contains "$output" "📋 Job is still active; Kubernetes may start a replacement execution (backoffLimit permitting)"
 }
 
 @test "kill_job_execution: warns when job already completed after pod deletion" {
@@ -408,5 +408,5 @@ teardown() {
   run bash "$BATS_TEST_DIRNAME/../kill_job_execution"
 
   [ "$status" -eq 0 ]
-  assert_contains "$output" "⚠️  Job for pod not found (it may have already completed)"
+  assert_contains "$output" "⚠️  Parent job not found (it may have already completed)"
 }

From 0e69faef69cf004970ace2ef979429aac05f409a Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 14:56:05 -0300
Subject: [PATCH 6/7] fix: guard K8S_NAMESPACE default under set -u in
 scheduled task trigger

---
 scheduled_task/scope/tests/trigger.bats | 39 +++++++++++++++++++++++++
 scheduled_task/scope/trigger            |  2 +-
 2 files changed, 40 insertions(+), 1 deletion(-)

diff --git a/scheduled_task/scope/tests/trigger.bats b/scheduled_task/scope/tests/trigger.bats
index 92e33fad..6ffd2576 100644
--- a/scheduled_task/scope/tests/trigger.bats
+++ b/scheduled_task/scope/tests/trigger.bats
@@ -177,3 +177,42 @@ teardown() {
   [ "$status" -eq 0 ]
   assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
+
+# The trigger-job workflow does not include values.yaml, so K8S_NAMESPACE is
+# unset at runtime. Under `set -u` an unguarded expansion aborts the script;
+# these reproduce that real environment.
+@test "trigger: does not abort under set -u when K8S_NAMESPACE is unset (uses provider namespace)" {
+  unset K8S_NAMESPACE
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
+}
+
+@test "trigger: falls back to the nullplatform default when K8S_NAMESPACE is unset and the provider namespace is absent" {
+  unset K8S_NAMESPACE
+  export CONTEXT=$(echo "$CONTEXT" | jq 'del(.providers["container-orchestration"].cluster.namespace)')
+
+  kubectl() {
+    case "$*" in
+      "get cronjob -n nullplatform"*)
+        echo "my-cronjob"
+        return 0
+        ;;
+      "create job --from=cronjob/my-cronjob"*"-n nullplatform")
+        return 0
+        ;;
+      *)
+        echo "unexpected namespace: $*" >&2
+        return 1
+        ;;
+    esac
+  }
+  export -f kubectl
+
+  run bash "$BATS_TEST_DIRNAME/../trigger"
+
+  [ "$status" -eq 0 ]
+  assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
+}
diff --git a/scheduled_task/scope/trigger b/scheduled_task/scope/trigger
index 7a2f47d7..0aeefc5a 100644
--- a/scheduled_task/scope/trigger
+++ b/scheduled_task/scope/trigger
@@ -16,7 +16,7 @@ if [[ -z "$DEPLOYMENT_ID" ]]; then
   exit 1
 fi
 
-K8S_NAMESPACE=$(echo "$CONTEXT" | jq -r --arg default "$K8S_NAMESPACE" '
+K8S_NAMESPACE=$(echo "$CONTEXT" | jq -r --arg default "${K8S_NAMESPACE:-nullplatform}" '
   .providers["container-orchestration"].cluster.namespace // $default
 ')
 

From e126c21ae01a86cda42a23dfc915a2bbf1d48fbe Mon Sep 17 00:00:00 2001
From: Federico Maleh <fedemaleh@yahoo.com.ar>
Date: Tue, 14 Jul 2026 15:38:32 -0300
Subject: [PATCH 7/7] fix: include values.yaml in trigger-job workflow to set
 K8S_NAMESPACE

---
 scheduled_task/scope/tests/trigger.bats       | 24 ++++++++++++++++---
 .../scope/workflows/trigger-job.yaml          |  2 ++
 2 files changed, 23 insertions(+), 3 deletions(-)

diff --git a/scheduled_task/scope/tests/trigger.bats b/scheduled_task/scope/tests/trigger.bats
index 6ffd2576..18ff2732 100644
--- a/scheduled_task/scope/tests/trigger.bats
+++ b/scheduled_task/scope/tests/trigger.bats
@@ -178,9 +178,9 @@ teardown() {
   assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
 
-# The trigger-job workflow does not include values.yaml, so K8S_NAMESPACE is
-# unset at runtime. Under `set -u` an unguarded expansion aborts the script;
-# these reproduce that real environment.
+# The workflow includes values.yaml to set K8S_NAMESPACE, but the script also
+# guards the expansion so that a missing include can never abort it under
+# `set -u`. These reproduce that unguarded environment (K8S_NAMESPACE unset).
 @test "trigger: does not abort under set -u when K8S_NAMESPACE is unset (uses provider namespace)" {
   unset K8S_NAMESPACE
 
@@ -216,3 +216,21 @@ teardown() {
   [ "$status" -eq 0 ]
   assert_contains "$output" "✅ The job my-cronjob was triggered, you can follow the execution from the logs screen"
 }
+
+# =============================================================================
+# Workflow wiring (trigger-job.yaml) — the layer the script tests can't see
+# =============================================================================
+@test "trigger-job workflow includes values.yaml so K8S_NAMESPACE is provided" {
+  run grep -A2 "^include:" "$BATS_TEST_DIRNAME/../workflows/trigger-job.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "values.yaml"
+}
+
+@test "trigger-job workflow loads the log function before the trigger step" {
+  run cat "$BATS_TEST_DIRNAME/../workflows/trigger-job.yaml"
+
+  assert_equal "$status" "0"
+  assert_contains "$output" "name: load logging"
+  assert_contains "$output" "\$OVERRIDES_PATH/logging"
+}
diff --git a/scheduled_task/scope/workflows/trigger-job.yaml b/scheduled_task/scope/workflows/trigger-job.yaml
index 954bccbd..02df28f0 100644
--- a/scheduled_task/scope/workflows/trigger-job.yaml
+++ b/scheduled_task/scope/workflows/trigger-job.yaml
@@ -1,3 +1,5 @@
+include:
+  - "$SERVICE_PATH/values.yaml"
 provider_categories:
   - container-orchestration
   - cloud-providers