For AI agents: The documentation index is at https://docs.digitalocean.com/llms.txt. Markdown versions of pages use the same URL with index.html.md in place of the HTML page (for example, append index.html.md to the directory path instead of opening the HTML document).

Spend alerts send email notifications when spending on your team or organization crosses a percentage threshold of a monthly budget you define. You can create multiple alerts that track total spend, specific products, or individual Serverless Inference model access keys, and organizations can scope an alert to a single team. Product-specific alerts let you track spending on individual products, such as Droplets or Kubernetes, separately from your overall total. Model access key alerts can also pause Serverless Inference requests for a key when its spend reaches its budget.

Spend alerts are available for both teams and organizations. For teams that do not belong to an organization, team members with the owner or biller role can create and edit alerts. For organizations, organization members with the owner or biller role can create and edit alerts, including alerts scoped to a single team in the organization.

To open the spend alerts page:

• For teams, go to the Spend alerts tab on the Billing page.
• For organizations, click the profile icon in the top right of the Control Panel, click View Organizations, select your organization, click the Billing tab, then click Spend alerts.

Create a Spend Alert

On the Spend alerts page, click Create alert to open the Create New Spend Alert page, then fill out the following fields:

• Alert name: A name for the alert.
• Scope (organizations only): Choose Organization total spend to track combined spend across all teams in the organization, or Specific team to limit the alert to one team’s spend.
• Monthly budget amount: The monthly spend amount that percentage thresholds are measured against.
• Alert thresholds: One or more percentages of your budget. We send a notification when your usage crosses each threshold. To use a percentage not in the list, click Add custom threshold.
• What do you want to track?: Choose one of the following options:
  • Total spend: Track all usage within the alert’s scope.
  • Specific products: Track only the products you choose. See Track Spending on Specific Products.
  • Managed Access Keys (Serverless Inference): In public preview, track spend for one or more Serverless Inference model access keys, and optionally pause a key when it reaches its budget. See Track Spending on Model Access Keys.

The Notification recipients section lists who receives the alert emails: team owners and billers for team alerts, and organization owners and billers for organization alerts.

Click Create alert to save the alert.

Track Spending on Specific Products

When you choose Specific products, a product list appears. Products are organized in three levels: groups, such as Compute and Managed Databases; product types within each group, such as Droplets and Kubernetes under Compute; and individual products within some product types, such as CPU Optimized under Droplets. Click a group or product type to see what it contains. Selecting a checkbox tracks everything under that item, and a partially filled checkbox means only some of the items under it are selected.

The list starts with all products selected. To clear every checkbox and start your selection from scratch, click Reset scope.

The alert measures the combined spend of the selected products against the monthly budget, and usage of other products does not count toward its thresholds. To get separate notifications for each product, create one alert per product, each with its own budget. For example, you can create one alert with a $500 budget for Droplets and another with a $200 budget for Kubernetes.

Track Spending on Model Access Keys public

Model access key alerts and pausing are in public preview. You acknowledge the preview when you enable pausing on an alert.

When you choose Managed Access Keys (Serverless Inference), the Choose Managed Access Keys to monitor list appears. Select up to 10 model access keys for the alert. For organizations, the list includes keys from the organization’s teams and shows which team generated each key. Legacy model access keys do not appear in the list and cannot have key alerts.

Spend is tracked per key. The monthly budget and thresholds apply to each selected key separately, not to the combined spend of all selected keys, and each key gets its own alert. A key can belong to only one alert, so keys that already have an alert cannot be selected again.

By default, a model access key alert only sends notifications, and Serverless Inference requests continue after the key’s spend reaches 100% of its budget. To stop requests instead, turn on Pause Serverless Inference at threshold, then select the checkbox to confirm that the selected keys are paused at 100% of the monthly budget. The setting applies to every key selected in the alert, and each key pauses independently when its own spend reaches the budget. Pausing affects only Serverless Inference requests authenticated with the paused key. For how pausing works and how paused keys resume, see Set Spend Thresholds on Keys.

Edit a Spend Alert

The Spend alerts page lists each alert’s name, scope, monthly budget, and thresholds. The list does not show what an alert tracks or whether it pauses keys. To check those settings, open the alert for editing.

To edit an existing alert, on the Spend alerts page, click the … menu next to the alert, then click Edit alert to open the Edit Spend Alert page. Update the alert’s name, budget, thresholds, or tracking options, including whether a model access key alert pauses the key, then click Save changes.

If a model access key alert has paused a key, raising the budget above the key’s spend, turning off Pause Serverless Inference at threshold, or removing the key from the alert resumes the key right away.

Delete a Spend Alert

To delete an alert, on the Spend alerts page, click the … menu next to the alert, then click Delete alert. In the Delete alert window, click Delete. You cannot undo the deletion.

Deleting a model access key alert resumes any keys it paused. Those keys have no spend threshold until you add them to another alert.

How Spend Alert Notifications Work

We send notifications based on your actual usage and notify recipients within an hour of your spend crossing a threshold. Each threshold notifies once per billing cycle. If your spend crosses more than one threshold between checks, we send one notification for the highest threshold crossed. If you change an alert’s budget or thresholds during a billing cycle, its threshold tracking resets, and notifications can fire again as spend crosses the updated thresholds.

The amount an alert measures depends on its scope:

• Team alerts and organization-wide alerts: Measure spend after credits and discounts are applied.
• Organization alerts scoped to a single team: Measure the team’s usage before credits and discounts, because the organization applies credits and discounts to its consolidated invoice rather than to individual teams.
• Model access key alerts: Measure the Serverless Inference spend of each key separately.

Because alerts are based on actual usage, projections of Droplet transfer in excess of the transfer pool do not trigger an alert until they are applied to the invoice.

Migrated Billing Alerts

Before spend alerts, you could set a single billing alert that emailed you when your monthly spending exceeded a set amount. If you had a billing alert, we migrated it to the Spend alerts page as an alert named Account Spend Alert with a single threshold at 100% of your previous alert amount. To receive notifications before your spending reaches that amount, edit the alert and add lower thresholds.