Sandbox

Try PingPulse in action - send a webhook ping and watch it update in real-time

How to Use This Sandbox

  1. Click the "Ping Stage" buttons to send webhook requests
  2. Watch the stage circles turn green as they receive pings
  3. Try pinging out of order (e.g., Stage 3 before Stage 2) to see the difference between modes
  4. Check the response log below to see status updates

Strict Mode

Strict Mode

Stages must execute in order. Out-of-order pings are rejected and missing stages are marked as MISSED.

1
Data Load
2
Processing
3
Validation
4
Complete
Response Log
No pings sent yet. Click a button above to start!

Flexible Mode

Flexible Mode

Stages can execute in any order. All pings are accepted regardless of sequence.

1
Data Load
2
Processing
3
Validation
4
Complete
Response Log
No pings sent yet. Click a button above to start!

Process Hooks

Simulate a long-running job by sending start and finish pings.

Uses sandbox key
Sandbox hook ID
Not created yet
Start URL
Finish URL
ML training
Idle
Time since start
Response Log
Click “Send Start” to begin.

Dynamic Workflow Script

Run a sample script that generates a dynamic workflow with branching, then renders the DAG below.

/dhk + tree → DAG
Sandbox workflow ID
Not run yet
What the script does
  • Creates a fresh sandbox workflow on first ping (start=yes)
  • Sends a sequence of stages with different agent_code
  • Runs a branch: 2.1 and 2.2 in parallel
  • Fetches /dhk/<id>/tree and renders the DAG
AI Agent Colors
orchestrator
data-agent
gpu-agent-a
gpu-agent-b
deploy-agent
Response Log
Click “Run Script” to generate a branched dynamic workflow.

Run the script to render the workflow graph…

Configuration Examples

See how PingPulse alerts and notifications are configured for your AI agent workflows.

Handoff Alerts

Stage transition monitoring

Alert when an AI agent fails to hand off to the next stage within expected time bounds.

Example: Agent Handoff Config
{
  "workflow_id": "dwf_ml_pipeline",
  "handoff_rules": [
    {
      "from_stage": "2",
      "to_stages": ["2.1", "2.2"],
      "timeout_seconds": 30,
      "alert_on_timeout": true
    },
    {
      "from_stage": "3",
      "to_stage": "4",
      "timeout_seconds": 60,
      "require_all_branches": true
    }
  ]
}
Handoff OK Pending Timeout
Detect stuck agents before pipeline stalls
Branch completion verification
Multi-agent coordination monitoring

Notification Channels

Alert delivery configuration

Configure where and how alerts are delivered when issues are detected.

Email
All plans
Webhook
Startup+
Slack
Startup+
Discord
Startup+
Example: Alert Notification Config
{
  "channel_type": "slack",
  "webhook_url": "https://hooks.slack.com/...",
  "events": [
    "stage_timeout",
    "handoff_failed",
    "workflow_completed",
    "agent_error"
  ],
  "filters": {
    "severity": ["warning", "critical"],
    "workflows": ["dwf_ml_*"]
  }
}
Supported Alert Events:
stage_timeout handoff_failed workflow_completed agent_error branch_merged heartbeat_missed

What's Happening Behind the Scenes?

When you click "Ping Stage", a simple HTTP GET request is sent to:

PingPulse:

  • Queues the incoming webhook immediately
  • Creates or updates the workflow instance
  • Validates the stage execution (strict/flexible mode)
  • Records the execution with timestamp
  • Logs to audit trail with source IP
  • Returns a JSON response with status

All of this happens in milliseconds, and you can monitor it in real-time in the dashboard!

Ready to Get Started?

Create your own workflows and start monitoring your pipelines.