Skip to main content

About Cloud Watcher

As a cloud-native product, Edge makes it simple and easy to run applications, perform analysis in Jupyter notebooks, and manage your data. But as many organizations have discovered, the cloud's ease-of-use also means "ease-of-spending." In this era of carefully-managed budgets, it's more essential than ever to provide complete control over the costs associated with cloud usage.

That's where Cloud Watcher comes in. Cloud Watcher is a suite of tools for monitoring and controlling your cloud spend with Edge. Importantly, Edge isn't limited to passive monitoring and sending reports to a system administrator. Costs are aggregated per-organization and made visible to any users you choose to authorize; for example, line managers. If a budget is exceeded, Edge can take action to "put the brakes" on cloud spending; for example, by disabling new application launches until a manager can investigate.

The Edge approach to cost control

Edge is based around three fundamental principles for controlling cost:

  1. Built-in architectural cost optimization, paying for only what you use. The best way to control costs is to avoid spending by default. Because Edge applications are purely pay-as-you-go, when not being used they incur zero compute costs. Expensive GPU instances can be spun up for only the amount of time needed to answer a question or perform an analysis.

  2. Total visibility and budget planning, for managers and not just server administrators. 100% of your costs are visible on the Cloud Watcher dashboard, not only to server admins but on a per-organization basis. Budgets allow you to set a maximum spending amount by month, quarter, or a custom time frame. If a budget is exceeded, Edge will send a notification right away, and even take action to limit further spending.

  3. Finally, allowing fine-grained control over functionality that can incur cloud costs. For example, you can customize which compute profiles are available to users in an organization; not everyone needs access to useful but high-priced compute resources. A common optimization for app launch time is to have a few machines running on standby; Cloud Watcher allows you to control how many of these "placeholders" are running and when they shut down. The system is designed to put you back in control.

Next Steps

See the next section, Tracking and Viewing Costs for an introduction to the basic functionality of Cloud Watcher.