If you have a large team using google cloud platform resources, you may want to make sure you set some limits on how much they can spend on certain types of resources. In this video we'll talk about what quotas are and how to set them up. Budgets allow you to trigger notifications, so that you can know how your actual and forecasted costs are trending over time and prompt you to take action. Quotas on the other hand, provide an extra advantage. They allow you to set a limit on the number of concurrent resources in a project or the number of API requests. Quotas are important to protect the community of GCP users and safeguard against unforeseen spikes in usage that could lead to budget overruns. There are two types of quotas, allocation and rate. Allocation quotas measure and limit a resource like the number of virtual machines you can have spun up at once. And rate quotas reset after a certain amount of time like API requests per minute. Most resources and APIs in each project have quotas set by default and some can even be unlimited. But you can also configure specific quotas to help manage your cloud resources and establish cost controls. For example, let's say you have a test environment where a lot of users have access to big query. Since it's a test environment, people maybe a little less careful about optimizing their queries which may result in some costly mistakes. So you may want to set up a quota limit to ensure each user can only query a certain amount of data each day. Let's look at how to do this by viewing and modifying an existing quarter limit. In the GCP console, you can use the top left navigation to select quotas under IAM and admin. Since quotas are set per project, you may need to choose a project if you haven't selected one already. You'll also need to have the right permissions for the projects such as project owner or editor in order to change any quota limits. This page contains information about all the quotas for GCB products and services available to you, there's a lot. So you can use the filters to narrow down the quotas you're interested in. Filtering by type lets you choose between unlimited quotas or quotas with usage. Service metric and location filter down to just quotas matching what you select. For our example, we'll start by filtering to just bit query to keep the list manageable. You can click on the service drop down, click on none to deselect everything, and then click big query API. Now we're looking at just the big query related quotas. Each row is a different quota, and you can click on a quota to go to the API specific quota page which may show you graphs or other details. On our current project wide quota page, each roll starts with a service quota description, then the location or regions that the quota affects. After that, the current usage and peak usage over the past seven days is shown. For allocation quotas, the current usage is how much of that resource is currently in use, while for rate quotas current usage refers to how much has been used so far in the current day. The last column is the actual limit enforced by the quota. The first item on the list is query research per day, which shows us how much data we've created today and that there is currently no limit set up. If we edited this quota, we could set a limit on the amount of data that could be queried per day for the entire project. For our example, we want to limit each user on the project. So, let's select the query usage per day per user quota instead. We'll select the checkbox, and then hit edit quotas at the top of the page. It's also possible to edit multiple quotas at once. On the right side of the console, a formal open up where we can input our limit. Some edits may require you to fill out contact information and a form to request the limit update, rather than modifying the limits directly. Quota requests are typically handled within two business days, and an email confirmation would be sent out once the requested update has been reviewed. For this quota, we can simply lower it from unlimited to any number we want without needing to make a request. Let's limit each user to half a terabyte of data queried. With the quota limits set, there will now be a limit for each user when they run queries. One important note is that the quota page doesn't show you the current usage for every quota, such as the per user quota we just edited. If you want to learn more about monitoring quotas, check out the technical documentation at cloud.google.com/docs, then search for monitoring quota metrics. To make sure we've set up the limit correctly, we can run a query in big query that is projected to process much more data than what we've allowed. If it's set up correctly, big query will display this error that shows us that the quota has exceeded and the query won't run. We could also run smaller queries up until our quota limit, at which point we can't run anymore until the next day. Now you should be able to view your existing quotas and update their limits to protect against any surprises. We recommend setting up quota limits that best match your desired usage across many GCP services. It's also possible to set up monitoring for quota metrics in stack driver, so you could create custom dashboards or alerts. Find out more in upcoming videos
