Documentation

Storage Buckets

Storage Buckets provide a unified, persistent storage layer natively integrated into your SolutionEngine project. Unlike environments which handle execution, Buckets act as the central repository for your data, aggregating information regardless of where the processing occurs.

Whether your workflows run on Edge Agents in a factory or on a cloud cluster, they can read from and write to the same Project Buckets securely.

Core Capabilities

A single bucket is capable of handling three distinct types of data simultaneously.

1. Metadata Storage (Key-Value)

Metadata storage allows you to save and retrieve unstructured JSON documents.

  • How it works: You define a "key", and store any JSON object against it.
  • Use case: Storing the current operational state of a piece of machinery, configuration overrides for a specific workflow, or maintaining a tally of objects detected across a shift.
  • Node Integration: Use the SaveMetadata node to seamlessly push dictionary payloads into the bucket. Use the ReadMetadata node to retrieve specific keys dynamically during a workflow.

2. Timeseries Data

Timeseries storage is optimized for timestamped metric data points.

  • How it works: You push key-value pairs (like {"temperature": 45.2}) into the bucket, and the system automatically appends a timestamp and indexes it chronologically.
  • Use case: Logging system health metrics (CPU usage, latency), tracking anomaly confidence scores over time, or recording event frequencies (e.g., "Person detected at 14:02").
  • Node Integration: Use the SaveTimeseries node. You can enable "Auto Timestamp" to log exactly when the node executed. Use ReadTimeseries downstream to fetch the "Latest Values" or filter data over a specific "Custom Period" (e.g., the last 24 hours).

3. Media & Asset Storage (Blob)

Media storage securely holds raw binary files, primarily used for images and video clips.

  • How it works: Binary data is saved into virtual folders called "Collections" within the bucket.
  • Use case: Archiving cropped images of defective products for future model retraining, or saving 10-second video clips immediately following a security event trigger.
  • Node Integration: Use the SaveMedia node to route a binary frame (e.g., data.image) from your workflow into a specific bucket collection. Use ReadMedia to retrieve files for auditing or secondary batch processing.

Workflow Integration

Buckets are intrinsically linked to the Workflow engine. You do not need to manage API keys or authentication when a workflow interacts with a bucket in the same project.

In the Visual Editor, look for the Storage / Buckets category in the Node Palette.

Available Storage Nodes:

  • SaveMedia / ReadMedia
  • SaveMetadata / ReadMetadata
  • SaveTimeseries / ReadTimeseries

When configuring these nodes, you simply select the target bucket from a dropdown menu. The backend handles the secure routing of the data.

Managing Buckets

You can manage your buckets directly from the Project Dashboard -> Buckets tab.

  • Creation: Click "Create Bucket" and provide a distinct name (e.g., manufacturing-line-a-metrics).
  • Exploration: Clicking on a bucket opens its detailed view, where you can tab between Media, Metadata, and Timeseries data to manually inspect the stored information.
  • Security: Buckets inherit the security boundaries of their parent Project. Users with "Viewer" access to a project can read bucket data, but cannot delete the bucket itself.
PreviousTransformation Nodes
NextModel Library