Overview of Scaling Resources and Performance Optimization on FME Flow Hosted

Introduction

There are two main things that can be altered on FME Flow Hosted (formerly FME Cloud) to control the performance, the size of the instance FME Flow (formerly FME Server) is running on, and the speed and size of the disk.
 

Tailoring Read/Write Disk Speed

If you are reading/writing large amounts of data on FME Flow, the read/write speed of the disk may be a bottleneck. FME Flow Hosted runs on AWS EC2 instances and utilizes AWS General Purpose SSD (GP2) Volumes. The important thing to know about this is, in general, the larger the disk, the higher the minimum input/output operations per second (IOPS). That means you may need to over-provision a disk if you are looking for extremely fast read/write speeds.

More information on sizing your disk correctly on FME Flow Hosted is available here.
 

Modifying the Instance Size

For any instances you are running, you can always increase or decrease the size of it. You can do this both in the Web UI and via the FME Flow Hosted API. When you change the size of your instance, FME Flow is restarted, so the instance will be offline for this duration (usually a couple of minutes). No data will be lost during the resizing, and the new instance will be an exact copy of your old FME Flow.
 

Scaling Compute Up and Down

You may not wish for your FME Flow instance to run all of the time on FME Flow Hosted. You can provision FME Flow capacity on FME Flow Hosted in an automated way. There are two main scenarios:

1. Provision an FME Flow to run only when required to keep costs down.
2. Add FME Flow capacity to an existing deployment dynamically.

Schedule Capacity

Built directly into FME Flow Hosted is a scheduler that enables you to schedule an instance to start and stop at a set time. For example, you might want a Standard instance to come online between 9:00 am and 5:00 pm, Monday to Friday.

More information on creating a schedule on FME Flow Hosted is available here.
 

Using the FME Flow Hosted API

The FME Flow Hosted API allows you to provision FME Flow capacity dynamically. You can launch, pause, and start instances programmatically. The most popular workflow is to launch an FME Flow manually, fully configure the FME Flow with your integration jobs, and then use the API to automatically start and pause the instance as needed, for example, to run FME workflows in response to an event. This is a powerful workflow as it ensures that you only pay for FME Flow when used.

Based upon a load of an external system, you can also vary the capacity on FME Flow Hosted This blog post walks you through such a scenario using AWS Lambda and the FME Flow Hosted API.