Monitoring Built for AWS

We auto scale with you, reacting instantly whenever your environment changes, with no agents to install or shell scripts to cURL

Steve Boak
The Opsee Blog
Published in
3 min readMar 31, 2016


From our earliest days as a company we wanted the experience of using Opsee to be special. Monitoring isn’t an experience you’d normally call effortless or minimal, but it’s what we set out to achieve. With public cloud services like AWS, there’s a lot we can do to make monitoring effortless. We have a decent idea what your services and systems are doing, and what they’re supposed to be doing, using environment data. That’s the basis for the experience we’re trying to create.

Getting Started

All you need are AWS keys to get started with Opsee. We don’t install agents on your systems–instead, we add a new EC2 instance to your environment that’s responsible for health checking and discovery:

Just enter AWS keys. We’ll add our instance and cover your environment

Here’s how it works:

  1. You use our IAM guide to create a temporary user for installation. Importantly, these keys are used only once and never stored. You can delete the user as soon as installation is done
  2. You tell us where to add our instance. Just choose an AWS region, VPC, and subnet and we’ll do the rest. We use a CloudFormation template and AMI for our instance, and its permissions are defined by a couple of IAM roles

That’s it! Once our instance is added, and you can easily create health checks for your infrastructure and services–we even try to create checks for you. We do a lot more than help you get started with monitoring, though. Our deep integration with your AWS environment is running continuously.

Rich Context and Actionable Notifications

Instance actions like restarts are simple

Two of the biggest benefits to using Opsee are rich and actionable notifications. Since we know where a failure occurs within a target group, we:

  1. show you the failing response in our notifications
  2. point you to the failing instance(s) in a group, and
  3. give you quick access to instance actions like restarts in Opsee.

Security Benefits

There are some notable security benefits to this model:

  • Since we don’t install any software on your systems, the probability of us crashing them is significantly lower than it would be with an agent
  • Our instance credentials are cycled every hour, making them a lot harder to grab
  • Any credentials required to run health checks are stored in an encrypted EBS volume inside your environment
  • Our instance’s auto scale group ensures that you always have one running
  • Uninstalling is as easy as activating. Should you ever need to, you can completely remove Opsee by simply deleting our CloudFormation stacks

Discovery and Auto Scaling

The most important thing to know about Opsee health checks is that when they’re attached to dynamic targets they never need updating. Discovery is running continuously, adapting when your environment changes. Our instance captures environment data using AWS APIs, keeping your checks up to date. There are no configurations to manage or YAML files to update.

Plans for the Future

This model lets us make some pretty interesting updates in the future. For example:

  • Running multiple, redundant instances inside your environment can make Opsee highly available, even in parts of your network that aren’t exposed to the internet
  • Opsee could run without a persistent connection to the internet, allowing it to continue running checks offline. This is interesting territory for a SaaS monitoring product

We think this is a better way to do monitoring: deeply integrated with your environment, with richer context, better performance, and a lot less work to get started and maintain.



Steve Boak
The Opsee Blog

VP of Product Design @datadog. Try my product.