Skip to content
The Things Stack documentation home
Getting Started Concepts Cloud Enterprise Integrations API Hardware
Getting Started Concepts
Cloud Enterprise
Integrations API Hardware
Get The Things Stack

    Recent searches

      Popular pages

      • Getting Started
      • Adding Gateways
      • Adding Devices
      • Payload Formatters
      • Webhooks
      • Authentication

      No results for “”.

      • Check the spelling, or try fewer or more general words.
      • Ask the community on The Things Network Forum
      • Contact support
      ↑↓ to navigate ↵ to open Esc to close

      The Things Stack Enterprise

      • Overview
      • AWS
        • AWS Marketplace AMI
        • AWS ECS
          • Architecture
          • Prerequisites
          • Deployment
          • Database Operations
          • Monitoring
          • Updating
          • Best Practices
          • Interoperability
          • Mutual TLS
          • Troubleshooting AWS ECS Deployment
          • Template Changelog
      • Kubernetes
      • Docker
      • The Things Join Server
      • Manage The Things Stack

      Best Practices

      Updated 20 January 2025 Edit on GitHub
      Copy page
      • View as MarkdownThis page as plain text
      • Open in ChatGPTAsk questions about this page
      • Open in ClaudeAsk questions about this page

      This page describes best practices of a The Things Stack deployment on AWS ECS.

      Stack Naming

      Templates are numbered in suggested order of deployment. This doesn’t mean that numbering won’t change in the future. Our suggested naming scheme is $PREFIX-$NAME, where $PREFIX is some prefix specific to your deployment.

      Staging

      Performing maintenance operations requires in-depth knowledge of AWS and The Things Stack. It is likely that certain action might have unexpected results. For this reason, we strongly advise to maintain a staging environment, and test all updates on staging before deploying to production.

      High Availability

      Most components of The Things Stack offer some form of high availability - multiple ECS tasks, master/replica database nodes, etc. It is vital to make use of these functionalities in case of a random failure.

      Back-ups

      The Things Stack uses five types of storage:

      1. AWS RDS Aurora database

      2. AWS ElastiCache database

      3. TimescaleDB

      4. AWS EFS drive for monitoring

      5. S3 buckets

      AWS RDS Aurora and AWS ElastiCache databases keep important data. By default they have automatic backups. It is advised to periodically check that the backups are still performed correctly, and they restore successfully.

      TimescaleDB keeps storage integration data. Currently there is no backup, because this data is not considered important.

      AWS EFS drive for monitoring keeps historic monitoring data. This data isn’t considered important.

      You might want to consider turning on versioning on the S3 buckets. Currently this can’t be done from our templates, but you can do it in bucket settings. Data in these buckets is not vital for functioning of The Things Stack, but loss might result in incovenience.

      SSH access

      EC2 machines included in templates can be accessed via SSH. Our suggested access models:

      1. The EC2 machines can be configured to work with AWS Session Manager. Using AWS Session Manager allows administrators to access the machines without using SSH keys, and manage rights via AWS IAM.
      2. The templates offer a firewall setting restricting SSH access to a given subnet. This can be used to only allow access from internal network, or from an SSH gateway.

      Updating software

      It is advised to periodically perform sudo yum update on the machines to update software other than The Things Stack.

      Alerting

      The Things Stack supports integration with services like Opsgenie for sending alerts about basic diagnostics, like CPU/Memory usage or high number of warnings. It is crucial to make use of this alerting system.

      Multi-tenancy

      When purchasing an Enterprise license, consider setting up a multi-tenant environment from the beginning if you plan to add tenants later. Switching from a single-tenant to a multi-tenant environment requires a database migration.

      RDS Auto minor version upgrade

      Enabling the Auto minor version upgrade option on AWS RDS results in automatic upgrades of RDS instances to the latest minor version, subject to AWS marking them as auto-upgradeable. These upgrades may entail minor downtime. For precise control over downtime, it’s advisable to opt for manual upgrades and disable this option.

      ← Updating Interoperability →

      On this page

      • Stack Naming
      • Staging
      • High Availability
      • Back-ups
      • SSH access
      • Updating software
      • Alerting
      • Multi-tenancy
      • RDS Auto minor version upgrade

      Sections

      Getting Started

      Concepts

      Cloud

      Enterprise

      Integrations

      API

      Hardware

      Sitemap

      View our Sitemap

      Contributing

      GitHub

      About Us

      The Things Industries

      Privacy Policy

      Cookie settings

      Cookie Settings

      We use cookies to enhance your experience, analyze site traffic and deliver personalized content. Read our Privacy Policy.

      Cookie categories