Skip to main content

API Service

The TagoIO API service is how services, applications, and devices reach your project's data and features. Every operation depends on it, so the rest of the platform needs an active API instance to work.

This page sets the machine size and autoscaling for incoming API requests. API usage is tied to your billing costs, since compute, memory, and network use drive your charges. See the Bills page for how that adds up.

Open the page from the TagoIO & API area, under SERVICES in the sidebar.

Instance Settings​

The Instance settings section configures machine type, scaling, and cooldown parameters:

  • Machine: the machine type for each instance. Default is 1 vCPU / 2GB RAM.
  • Minimum instances: the lowest number of instances kept running. Default is 1.
  • Maximum instances: the highest number autoscaling can reach. Default is 1.
  • Scale on CPU utilization: the CPU percentage that triggers scaling. Default is 60.
  • Cooldown for scaling up: seconds to wait before adding instances again. Default is 200.
  • Cooldown for scaling down: seconds to wait before removing instances. Default is 300.

Click Save to apply changes.

Scaling Strategies​

As the project grows with more devices, users, and automated workflows, the API service has to handle more load. You scale it by adjusting the vertical config, the horizontal config, or both:

  • Vertical scaling (Machine): raise the CPU and RAM of each instance with a larger machine type. This helps with intensive data processing or large retrievals, such as dashboard queries returning more than 10,000 records in a single request. For machine types beyond the listed options, contact TagoIO support.
  • Horizontal scaling (instances): raise the minimum and maximum instance counts so requests spread across more instances. This is the better fit for high concurrency and traffic bursts, and it improves fault tolerance.

Autoscaling adds and removes instances between the minimum and maximum based on the CPU threshold and the two cooldown windows.

Monitoring​

The Monitoring section has a range toggle across 1h, 6h, 12h, 1d, 3d, 7d, and 30d. Charts show "No data available" until there is data for the window.

  • CPU Utilization (%): how intensively the API uses processing resources. Use it to spot peak periods and the operations driving the highest load.
  • Memory Utilization (%): how much data the API holds in memory while serving requests. Use it to catch requests or concurrent operations that push memory high, then tune data handling.

Billing Considerations​

The API runs on AWS, and resource use drives billing, so review the instance settings and the monitoring charts regularly. Tighter scaling settings cut wasted cost while keeping the project responsive.

For billing models, usage analysis, and cost details, see the Bills page.