Automations (Technical Reference)

This document covers the technical internals of RocketLead’s automation engine — how flows execute, what each node type does, and how variables are threaded between steps.

Execution Model

Flow Lifecycle

1. Trigger event — something happens (lead added, status changed, appointment booked, etc.)
2. Flow matching — the system finds all enabled, published flows with a matching trigger
3. Run creation — a FlowRun is created with status scheduled and linked to the published VersionedFlow
4. Step execution — each node is executed in order, following edges. Steps are processed via Cloud Tasks, allowing delays and retries.
5. Completion — when all reachable nodes have executed, the run is marked completed (or failed if any step fails)

Run Statuses

Status	Meaning
scheduled	Run created, first step queued
running	At least one step is executing
completed	All steps finished successfully
failed	A step failed without recovery
retrying	A step is being retried
interrupted	Run was cancelled while a step was pending
cancelled	Run was manually cancelled

Step Processing

Each step goes through:

1. Load the step and its parent run
2. Check if the run is cancelled/interrupted — if so, mark step as interrupted
3. Mark step as running
4. Resolve the node type and load its handler
5. Replace placeholder variables in node fields (see Variables below)
6. Execute the handler
7. Evaluate outgoing edge conditions
8. Queue the next step(s) via Cloud Tasks

Flow Chain Depth

Flows can trigger other flows (e.g., a lead status change triggers a second automation). The flowChainDepth tracks nesting to prevent infinite loops.

Node Type Reference

Triggers

Triggers start a flow. Each flow has exactly one trigger node.

Type	Label	Fires When	Config Fields
table.entry.added	Neuer Lead hinzugefügt	A new lead is added to a table (via webhook, form, or API)	tableIds[], execution timing
table.entry.leadState.changed	Leadstatus geändert	A lead’s status changes	tableIds[], oldLeadStateId, newLeadStateId, triggerOnCreation, execution timing
calendar.appointment.booked	Neuer Termin gebucht	An appointment is booked	calendarIds[], originTableId, execution timing
calendar.appointment.hasStarted	Termin beginnt	An appointment’s start time is reached	calendarIds[], originTableId, execution timing (supports negative = before event)
calendar.appointment.hasEnded	Termin endet	An appointment’s end time is reached	calendarIds[], originTableId, execution timing (supports negative)
calendar.appointment.deleted	Termin gelöscht	An appointment is deleted	calendarIds[], originTableId

Execution timing allows offsetting when the flow runs relative to the event (e.g., “1 day before appointment starts”).

Communication

Type	Label	What It Does	Config Fields
email.send	E-Mail senden	Sends an email via configured provider	from (email selector), to (comma-separated, supports variables), subject, body (rich text)
sendgrid.template.send	Sendgrid Vorlage versenden	Sends a SendGrid template email	from (sendgrid account), fromName, to, templateId, payload (JSON), attachmentUrl
sms.send	SMS senden	Sends an SMS via Reach	identityId, to (phone, supports variables), message

Actions

Type	Label	What It Does	Config Fields
followUpDate.set	Follow-Up terminieren	Sets a follow-up date on a lead	tableEntryId, referenceDate, offset (value + unit + direction)
webhook.post	POST Request	POSTs JSON to a URL	url, body (JSON)
table.entry.leadState.set	Leadstatus setzen	Changes a lead’s status	tableEntryId, tableId, leadStateId
httpRequest	HTTP Request	Sends a configurable HTTP request	method (GET/POST/PUT/PATCH/DELETE/HEAD/OPTIONS), url, headers (JSON), body, rawTimestamps (boolean)

Helpers

Type	Label	What It Does	Config Fields
variable.set	Variable setzen	Defines variables for downstream nodes	Key-value pairs (min 1)
wait	Warte	Delays execution	duration (value + unit: minutes/hours/days)
router	Router	Branches the flow based on edge conditions	No config fields — conditions are on the edges

Variables

Variables let you pass data between nodes. They use the {{placeholder}} syntax in text fields.

Variable Sources

1. Trigger context — data from the triggering event:

   • {{tableEntry.data.firstName}} — lead data fields
   • {{tableEntry.customFieldData.<id>}} — custom field values
   • {{booking.startDateTime}} — booking details

2. Studio variables — always available:

   • $studio.name, $studio.address, $studio.zip, $studio.city, $studio.country
   • $studio.phone, $studio.email, $studio.website

3. Global variables — set by variable.set nodes:

   • $global.<key> — accessible by all downstream nodes

4. Node output variables — output from upstream nodes (e.g., HTTP response data)

Variable Resolution

Before each step executes, all {{placeholder}} strings in the node’s config fields are replaced with actual values from the run context. Variables that can’t be resolved are left as-is.

Edge Conditions

Connections between nodes can have conditions that control flow:

Operator	Label	Description
equals	ist gleich	Exact match
not_equals	ist nicht gleich	No match
contains	enthält	Text contains value
not_contains	enthält nicht	Text doesn’t contain value
starts_with	beginnt mit	Text starts with value
ends_with	endet mit	Text ends with value
is_empty	ist leer	Field is empty or undefined
is_not_empty	ist nicht leer	Field has a value
greater_than	größer als	Numeric comparison
less_than	kleiner als	Numeric comparison

When a node completes, all outgoing edges are evaluated. Only edges whose conditions pass (or edges with no condition) lead to the next node being executed.

Data Model

Entity	Description
Flow	The automation definition: id, name, enabled, status
FlowNode	A block in the flow: type, fields (config), position (x/y)
FlowEdge	A connection: sourceNodeId → targetNodeId, optional condition and label
VersionedFlow	An immutable snapshot created on publish, with versionNumber
FlowRun	An execution instance linked to a VersionedFlow, with status, context, executeAt, completedAt
FlowRunStep	Per-node execution record with status, executeAt, completedAt, cloudTaskId