Chapter 56. Issues and Recommendations
Table of Contents
PPEM provides the issues and recommendations functionality, which monitors the instance metrics, detects issues, and suggests ways to resolve them. This functionality can be run automatically on a schedule or manually on demand.
Detected issues may include long-running transactions, long lock waits, deadlocks, autovacuum problems, replication lag, data checksum failures, and others. Each issue is associated with a metric, and each metric belongs to a certain category and has an importance level — Critical, High, Medium, or Low.
An issue is detected when the value of a monitored metric does not satisfy the corresponding detection rule. For each metric, the check collects its value during the monitoring window and compares it against the detection rule: if the value falls outside the condition set by the rule, the rule is triggered and the issue is detected. The check analyzes the metrics that are collected by the agent, so no additional setup is required from the user. Automatic check runs follow a schedule predefined by the system, and detection rules cannot be modified or deleted by the user.
Using this functionality, administrators can examine large amounts of monitoring data and separate primary issues from secondary ones.