Skip to main content

Main Database

The Main Database stores and maintains all your TagoIO data, including profile configurations and device data. If it becomes unavailable, the API cannot retrieve stored data, which affects every project feature.

This page configures the primary database. Open it from the TagoIO & API area, under SERVICES in the sidebar.

Database Settings​

The Database settings section configures machine type, read replicas, and backup parameters:

  • Machine: the machine type for the database. Default is 2 vCPU / 4GB RAM.
  • Additional reader instances: read-only instances added alongside the writer. Default is 0.
  • Backup retention period (days): how many days each backup is kept before deletion. Default is 7.
  • Daily backup schedule: the time of day, in UTC, when the daily backup runs. Default is 04:30.

Click Save to apply changes.

Backups​

Backups run automatically once a day at the scheduled time and are kept for the retention period, then deleted to manage storage. The schedule and retention are set here; existing snapshots are listed on the Backups page.

Scaling Your Database​

As demand grows, you may need to scale the database to handle more read and write operations:

  • Vertical scaling (Machine): move to a machine type with more CPU and RAM. This helps with large data submissions or high write and update activity.
  • Horizontal scaling (reader instances): add read-only instances to spread read-heavy operations and improve availability.

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. The following metrics are available:

CPU Utilization (%): Reads and writes consume CPU to process the data and run the database pipeline operations around them, such as validation, transaction management, and integrity checks. CPU can spike during heavy queries, like aggregates that calculate daily averages over a full month, and during updates that touch large entity or device datasets.

Freeable Memory (BYTES): the memory currently available to the database instance. By default the database caches frequently used queries, using around 90% of freeable memory, and evicts older entries to make room for newer ones. Consistently consuming all available memory usually means request volume or complexity exceeds what the cache can manage, and performance may degrade.

Database Connections (COUNT): the number of active connections to the database, including connections from API instances and other services.

Read IOPS and Write IOPS: the number of input and output operations per second. When investigating CPU or memory spikes, lining these up against IOPS at the same timestamps helps tell whether disk activity is driving the issue.

Network Throughput (B/S): the volume of data, in bytes per second, transmitted to and from the database.

For environments supporting up to 10,000 devices, this baseline is a good starting point:

  • Machine: 2 vCPU / 4GB RAM
  • Additional reader instances: 1
  • Backup retention period: 7 days