Anomaly Detection and Alerts

Alerts let you proactively monitor your integrations and respond before issues affect users. You can configure rules that trigger email notifications when latency, traffic, resource usage, or log activity crosses a defined threshold.

note

Alerts are configured at the integration level.

Alert types

WSO2 Cloud supports the following alert types:

• Latency alerts
• Traffic alerts
• Resource alerts
• Log alerts
• Build failure alerts

Latency alerts

Triggers when the response latency of an integration exceeds a threshold for a sustained period. Useful for integrations with SLA requirements or for proactively identifying slow integrations.

Parameter	Description
Metric	The latency percentile to monitor: 99th, 95th, 90th, or 50th.
Threshold	Latency in milliseconds (for example, 1800).
Period	Duration the threshold must be continuously exceeded (for example, 5 minutes).

Traffic alerts

Triggers when the request count exceeds a threshold. Useful for integrations with backend traffic limits or monetized backends that need proactive scaling.

Parameter	Description
Threshold	Requests per minute (for example, 200).
Period	Monitoring window (for example, 5 minutes).

Resource alerts

Triggers when CPU or memory usage exceeds a threshold. Use this to adjust resource allocations before performance degrades or downtime occurs.

Parameter	Description
Metric	CPU or Memory.
Threshold	In mCPU for CPU, or MiB for memory (for example, 1000).
Period	Duration the threshold must be continuously exceeded (for example, 5 minutes).

tip

• CPU: mCPU (milliCPU) measures CPU usage in fractions of a core, where 1000 mCPU = 1 full core.
• Memory: MiB (mebibyte) measures memory in binary units, where 1 MiB = 2²⁰ bytes.

Log alerts

Triggers when a specified phrase appears a minimum number of times in your integration logs within a defined time window. Useful for detecting recurring errors or critical events quickly.

Parameter	Description
Search Phrase	Keyword or phrase to monitor (for example, failed).
Count	Minimum number of occurrences to trigger the alert (for example, 10).
Interval	Time window for counting occurrences (for example, 5 minutes).

Build failure alerts

Triggers whenever a build failure occurs for your integration. Keeps your team informed so issues in the development workflow are caught early.

Configure an alert

info

You need DevOps or Platform Engineer permission to create alert rules.

1. Open the integration you want to monitor.

2. In the left navigation under Observability, click Alerts.

3. Click Create Alert Rule.

4. Select the Alert Type.

5. Select the Environment to monitor.

6. Select the Deployment Track or Version as required.

7. Configure the parameters specific to the selected alert type.

8. In the Emails field, enter the email addresses to notify when the alert triggers.

   note

   • Press Enter after typing each email address to add it.
   • You can add a maximum of 5 email addresses per alert.

9. Expand Advanced Configurations to set additional parameters if needed. Available options vary by alert type.

10. Review the Explanation callout to confirm the alert is configured as intended.

11. Click Create to save and activate the alert rule.

    info

    You can configure a maximum of 10 alerts per integration.

The alert appears in the Configure Alerts pane. From this pane you can edit, disable, enable, or remove any alert rule.

note

When editing an alert, Alert Type, Environment, and Deployment Track cannot be changed.

Alerts history and notifications

View alerts history

The Alerts History pane shows a record of alerts that have triggered for the integration. Filter by Alert Type, Environment, Deployment Track, Version, or Time Range to narrow the results. Click an alert entry to expand it and see full details.

Email notifications

When an alert triggers, all configured recipients receive an email with the alert details and a link to view it in WSO2 Cloud.