Skip to main content

Retool-managed deployment health monitoring and scaling

Learn about monitoring and scaling Retool-managed, self-hosted deployments.

Retool-managed, self-hosted deployments are available for invoiced customers. Contact your Retool account manager to learn more.

Retool monitors the health of a Retool-managed, self-hosted instance 24/7 and is automatically alerted of any potential issues that may arise. This coverage includes:

  • Containers (e.g., upgrade errors or misconfiguration)
  • Error log spikes (e.g., functionality degredation, capacity limits)
  • Databases (e.g., Amazon RDS storage or resource usage)

Health monitoring only applies to infrastructure of the Retool-managed, self-hosted instance. Refer to the observability documentation to learn how you can monitor object-level errors for apps, workflows, and more.

Alerts and response​

Retool immediately reviews any alert that is triggered. If the alert requires remediation, a Retool engineer begins to diagnose the issue. If necessary, the engineer then takes any action to resolve the issue that triggered the alert.

Alerts can be triggered for any number of reasons so not every alert requires remediation. In the event that an alert did require a Retool engineer to take action, you're notified of what occurred within one business day.

Observability data is configured for use only by Retool and cannot be shared with customers.

Deployment scaling​

Retool-managed, self-hosted deployments can scale as your needs grow. How it scales depends on its usage. You can contact Retool at any time to discuss scaling options. Retool will also proactively reach out if your deployment is reaching any limits.

RDS database size​

The RDS PostgreSQL database starts with 40GB of storage. It automatically increases in size as needed, up to the default limit of 200GB. Retool can raise the limit at any time upon request. There is no downtime whenever the database increases in size or the limit is raised.

The database only increases in size automatically. If the amount of data it contains is reduced, it does not scale down.

RDS instance type​

The RDS CPU and memory resources are determined by its instance type. The default type, db.m6g.large, uses 2x vCPUs and 8GiB memory. Unlike the RDS database size, CPU and memory resources do not scale automatically. Retool can upgrade—or downgrade—the instance type upon request. Retool also proactively monitors the instance and will recommend changes if they are deemed necessary.

Changes to the RDS instance type require downtime. They should only be performed during a regularly scheduled maintenance period.

Backend containers and resources​

Retool automatically scales the Retool-managed, self-hosted instance to ensure that it has sufficient capacity for the amount of traffic it needs. This includes making adjustments to the number and type of containers in use.

The default capacity for a Retool-managed, self-hosted instance is 20 queries per second (QPS). The resources required by your instance is dependent on your use case. If you require resources for a high volume of traffic, or if you expect an influx of traffic, contact your Retool account manager for guidance on scaling and preparation.