Getting Started
Quick Start
Get your first automation tracked in InfiniAnalytics in minutes. This guide walks you through creating a process, setting up an automation, understanding how events and executions work, and choosing an integration.
1. Create a Process
A Process represents a business workflow - for example “Monthly Invoice Processing” or “Customer Onboarding”. Processes group related automations under a shared department and hold the metadata (expected volume, time saved per run) that powers the ROI calculations in your analytics dashboard.
To create one, go to Processes in the sidebar and click New Process. Fill in:
- ●Name - A short, descriptive label for the business workflow.
- ●Department - The organizational unit that owns this process.
- ●Expected executions / year - Used to project annual ROI.
- ●Time saved per execution (min) - Drives the "hours recovered" KPI.
Save the process. It will appear in the catalogue and in your analytics views immediately.
2. Create an Automation
An Automation is the specific bot, script, or flow that carries out the steps within a process. Open the process you just created and click New Automation. Give it a meaningful name - e.g. “Extract Invoice Data” - and save.
You can create as many automations as you need inside a single process. Each one gets its own UUID, its own execution history, and its own performance charts.
3. Events
An Event is a timestamped signal your automation sends to InfiniAnalytics. Events are the building blocks of an execution timeline - they tell the platform what happened, when, and whether it succeeded or failed.
Every event carries at minimum: the automation UUID, an execution ID, an event type, and an optional message. Additional metadata (step name, custom payload) can be attached depending on the integration.
Event Types
There are five event types, and they form a lifecycle:
START
Marks the beginning of an execution. Exactly one START event should be sent per execution run. InfiniAnalytics uses this timestamp as the execution start time.
EVENT
An intermediate checkpoint inside an execution - e.g. "File downloaded", "Row processed", "API call completed". You can emit as many EVENT events as needed to capture progress milestones.
WARNING
Flags a non-critical issue worth reviewing - an unexpected value, a suboptimal config - without failing the execution.
END
Marks a successful completion. InfiniAnalytics uses this timestamp as the execution end time and records the execution as successful.
ERROR
Marks a failure. The execution is recorded as failed. You can send an ERROR at any point - in a try/except block, a flow error branch, or a global error handler.
START and ends with either END or ERROR. Events sent after an END or ERROR are accepted but will not change the execution outcome.4. Executions
An Execution is one complete run of an automation. InfiniAnalytics groups all events that share the same execution ID into a single execution record. The record stores the full event timeline, the start and end timestamps, the total duration, and the outcome (success or error).
You can browse all executions for a given automation on its detail page, or see executions across all automations from the Executions page in the sidebar. Clicking any execution opens the full event-by-event timeline.
The Execution ID
The execution ID is a string you generate on your side (typically a UUID v4) and include in every event you send for a given run. It acts as the correlation key - all events with the same execution ID are stitched together into one execution record.
This design gives you full control:
- ●You can generate it at the start of your script and pass it through every step.
- ●Multi-step workflows (e.g. a Power Automate flow that calls a Python subprocess) can share the same execution ID across different technologies.
- ●If your automation retries a failed step, you can choose to emit a new ERROR event under the same ID or start a fresh execution with a new ID - both patterns are valid.
5. Integrations
InfiniAnalytics works with the tools you already use. Pick an integration to continue:
- ▶Overview - Send events from any language or platform via HTTP.
- ▶Python SDK - Native library for scripts, agents, and backend services.
- ▶Overview - Track AI coding agents and LLM workflows via Model Context Protocol.
- ▶Power Automate Desktop - Ready-made action module for PAD flows and RPA bots.
- ▶Power Automate Cloud - HTTP connector for cloud flows and scheduled triggers.
- ▶n8n - Custom node for n8n self-hosted and cloud workflows.
