Skip to main content

In-Memory Database

The in-memory database in TagoIO aggregates all in-transit memory resources across multiple active API instances. This system is fundamental to real-time data processing, high-performance caching, and robust message queuing between APIs and microservices. By utilizing distributed in-memory queues, TagoIO prevents service overload and ensures that critical information is retained and processed, even during failures or traffic spikes. This architecture enables low-latency data delivery and high availability, supporting demanding enterprise use cases.

Key Use Cases​

Real-Time Data Transmission​

The in-memory database plays a critical role in real-time data delivery. For example, it enables instantaneous updates to dashboards as soon as new information becomes available, ensuring users always see the most current data without delay.

Caching System​

TagoIO implements a robust caching mechanism via the in-memory database to optimize performance for high-frequency requests. Common use cases include token authentication and other repetitive queries, where cached responses significantly reduce latency and backend load.

Scaling the In-Memory Database​

The in-memory database can be scaled to meet your project’s performance and reliability requirements, using two primary strategies:

  • Vertical scaling (Machine Tier): Involves increasing the CPU and memory resources allocated to each in-memory database instance. This approach is effective for handling extreme data ingestion peaks or supporting dashboards configured to display large volumes of real-time data.

  • Horizontal scaling (Number of machines): Adding reader replicas introduces additional read-only instances to the in-memory database cluster. This configuration distributes read operations, alleviating load on the primary instance and improving performance for applications with heavy read access patterns.

For most production environments, TagoIO recommends the following baseline settings:

  • Reader Replicas: 1 reader replica
  • Machine Specifications: 2 vCPU / 4GB RAM per instance

Monitoring and Metrics​

Effective monitoring of the in-memory database is essential for maintaining performance and optimizing resource allocation. Key metrics include:

Cache Hits​

This metric tracks the number of successful cache retrievals over time, indicating how effectively the cache is serving repeated requests and reducing backend workload.

Network Inbound/Outbound​

These metrics measure the volume of data, in bytes, transmitted to and from the in-memory database. Monitoring network throughput helps identify bottlenecks and ensures the system is sized appropriately for real-time and high-throughput use cases.

By understanding and configuring your in-memory database according to these guidelines, you can ensure robust, low-latency data delivery and high availability for all TagoIO services.