The Alerts screen lists alerts raised when a container breaks one of your rules, or when its usage departs sharply from its own recent pattern — such as a sudden spike or steadily climbing memory that looks like a leak. Checks run every minute.
Unusual-usage detection
You do not need any rules for this. PS-pocketEdge learns each container's normal usage and raises an alert when it changes sharply, including a "Possible memory leak" alert for memory that keeps growing.
Create a rule
- Open Alerts, go to Rules and click New rule.
- Name the rule, for example High CPU.
- Choose the metric — CPU, memory as a percentage of the container's limit, or process count — and the threshold it must stay above.
- Set how long it must stay above the threshold (up to 180 minutes) before alerting.
- Limit it to one server or a container name if you like. Container names match exactly and keep working when the container is recreated.
- Choose Warning or Critical, and save.
Handling alerts
Firing alerts show when they started. Click Acknowledge to show your team someone is on it. Alerts resolve on their own when the container is back within its rule; turn on Show resolved to see past alerts and how long they lasted. Deleting a rule resolves its open alerts but keeps them in the history.
Recommended limits
A container's Resources tab includes Insights. After about 30 minutes of usage history it compares the container's CPU, memory and process limits with what it actually uses. When they do not fit, it suggests new values; click Review & apply to use them.
Example: sustained CPU use
Create a Warning rule for CPU (cores), Above 0.9, For at least 5 minutes, scoped to your server and the exact container name demo-web. This tests consumption above 0.9 CPU cores, not 90% of the server’s total CPU. A brief spike does not satisfy the sustained duration. After the alert fires, acknowledge it while investigating; it resolves when usage returns within the rule.