Skip to content
Back to skills

2782 Actions 95a9b5cc

ASecurity

Complete reference for Robusta playbook actions.

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
devopspythonbashnodekubernetesgitapi

Works with

  • api

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 2782-actions_95a9b5cc --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 2782 Actions 95a9b5cc?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 2782 Actions 95a9b5cc
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-2782-actions-95a9b5cc/badge)](https://www.skillsdirectory.com/skills/tools-only-2782-actions-95a9b5cc)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
# Actions Reference

Complete reference for Robusta playbook actions.

## Log Actions

### logs_enricher

Adds container logs to alert.

```yaml
actions:
  - logs_enricher:
      container_name: main      # Optional: specific container
      previous_logs: true       # Include previous container logs
      warn_on_missing_label: true
      regex_replacer_patterns:  # Redact sensitive data
        - pattern: "password=.*"
          replacement: "password=REDACTED"
```

### pod_events_enricher

Adds Kubernetes events for the pod.

```yaml
actions:
  - pod_events_enricher:
      max_events: 10
```

## Graph/Metrics Actions

### pod_graph_enricher

Generates resource usage graphs for pod.

```yaml
actions:
  - pod_graph_enricher:
      resource_type: Memory    # Memory or CPU
      display_limits: true     # Show limits line
```

### cpu_graph_enricher

CPU-specific graph enricher.

```yaml
actions:
  - cpu_graph_enricher:
      resource_type: CPU
```

### node_cpu_enricher

Node-level CPU information.

```yaml
actions:
  - node_cpu_enricher: {}
```

### node_memory_enricher

Node-level memory information.

```yaml
actions:
  - node_memory_enricher: {}
```

### custom_graph_enricher

Custom PromQL query graph.

```yaml
actions:
  - custom_graph_enricher:
      graph_title: "Request Rate"
      promql_query: "rate(http_requests_total{pod='$pod'}[5m])"
      graph_duration_minutes: 60
```

### prometheus_enricher

Execute custom PromQL query.

```yaml
actions:
  - prometheus_enricher:
      query: "sum(rate(http_requests_total[5m]))"
```

## OOM Actions

### pod_oom_killer_enricher

Information about OOM kill events.

```yaml
actions:
  - pod_oom_killer_enricher: {}
```

### oom_killer_graph_enricher

Memory graph with OOM events marked.

```yaml
actions:
  - oom_killer_graph_enricher:
      delay_graph_s: 30  # Wait before generating
```

## Status Actions

### pod_status_enricher

Pod status and conditions.

```yaml
actions:
  - pod_status_enricher: {}
```

### deployment_status_enricher

Deployment status details.

```yaml
actions:
  - deployment_status_enricher: {}
```

### node_status_enricher

Node status and conditions.

```yaml
actions:
  - node_status_enricher: {}
```

### node_running_pods_enricher

List pods running on node.

```yaml
actions:
  - node_running_pods_enricher:
      max_pods: 20
```

### node_allocatable_resources_enricher

Node allocatable resources.

```yaml
actions:
  - node_allocatable_resources_enricher: {}
```

### cluster_status_enricher

Overall cluster status.

```yaml
actions:
  - cluster_status_enricher: {}
```

### related_pods_enricher

Find related pods (same deployment, etc.).

```yaml
actions:
  - related_pods_enricher: {}
```

## Remediation Actions

### delete_pod

Delete the affected pod.

```yaml
actions:
  - delete_pod:
      force: false          # Force delete
      grace_period: 30      # Seconds
```

### node_bash_enricher

Run bash command on node.

```yaml
actions:
  - node_bash_enricher:
      bash_command: "df -h"
```

### pod_bash_enricher

Run command in pod container.

```yaml
actions:
  - pod_bash_enricher:
      bash_command: "ls -la /app"
```

### restart_loop_reporter

Report restart loop with details.

```yaml
actions:
  - restart_loop_reporter:
      restart_reason: OOMKilled
```

## Report Actions

### event_report

Generate event report.

```yaml
actions:
  - event_report:
      title: "Kubernetes Event"
```

### create_finding

Create custom finding/alert.

```yaml
actions:
  - create_finding:
      title: "Custom Alert"
      aggregation_key: "custom-alert"
      severity: HIGH
      description: "Something happened"
```

## External Integration Actions

### send_to_webhook

Send data to external webhook.

```yaml
actions:
  - send_to_webhook:
      url: "https://my-api.com/alerts"
      headers:
        Authorization: "Bearer token"
```

### jira_issue_reporter

Create Jira ticket.

```yaml
actions:
  - jira_issue_reporter:
      project_key: "OPS"
      issue_type: "Bug"
```

## Silencing Actions

### silence_alert

Silence an alert temporarily.

```yaml
actions:
  - silence_alert:
      duration_minutes: 60
      comment: "Known issue, investigating"
```

## Customization

### Adding Custom Actions

Custom actions can be defined via Python:

```python
# In a custom playbook file
from robusta.api import *

@action
def my_custom_action(event: PodEvent):
    pod = event.get_pod()
    event.add_enrichment([
        MarkdownBlock(f"Pod {pod.metadata.name} info")
    ])
```

Load custom actions:

```yaml
# In generated_values.yaml
playbookRepos:
  - url: https://github.com/my-org/my-playbooks
    key: my-actions
```

## Action Parameters

All actions support these common parameters:

| Parameter | Type | Description |
|-----------|------|-------------|
| `rate_limit` | int | Seconds between executions |
| `timeout` | int | Action timeout in seconds |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.