The dashboard opens in a new tab when you click on Monitor > Dashboard. It shows critical information about the whole environment and is auto-refreshed every minute by default. The dashboard is ideal for displaying on a large TV screen.
The dashboard has four main areas: alerts, volumes, early warnings and query anomalies. The first one, alerts, is divided into current alerts, predictive alerts and CPU contention:
Current alerts shows the top 5 ongoing alerts: which node (server, instance or database) caused it, which performance counter it relates to, what the measured value and alert threshold are, and two gauges showing how often the alert happens and how many consecutive alerts there have been.
Predictive alerts is similar to current alerts, except that a time estimate for when the alert will happen is shown.
CPU contention is a situation where the processor has more waiting tasks in the processor queue than it is able to process in a decent time window. This causes CPU slowdown and can hurt overall performance. It typically happens on overloaded servers.
The CPU contention list shows any 4-minute time periods where
- The server's CPU usage was constantly at least 80%, and
- The server's CPU queue
- was at least 2 times the number of logical cores on a physical server, or
- was at least 3 times the number of logical cores on a virtual machine.
The alert shows the name of the server and the time when the contention started. It also shows the average CPU usage and processor queue length over the 4-minute period.
Volumes shows a top 5 list of server volumes with the least amount of free space left. To see the full list of volumes, click on the full report link.
Early warnings identifies workload performance counters which are statistically going into a worse direction. The change is calculated from the preceding month while the baseline is calculated from three months before. The following statistical components are inspected:
- Regression slope
- Workload volatility
- Skewness
- Kurtosis
Based on these components, the algorithm identifies performance counters which are at risk. We want to understand why the workload is headed into a worse direction before it starts to negatively affect the system performance.
As a rule of thumb, any red alert on a given performance counter should be clicked to see the details. The details view shows how each statistical component has evolved during the last month (actual) compared to the previous 3 months (baseline). The greater the deviation, the greater the change in the counter's workload profile.
If the actual values are above the baseline and the distance of the data points is growing exponentially, this may refer to a preliminary performance bottleneck, especially if this is true for more than one component (slope, volatility, skewness, kurtosis).
Query anomaly predictions show individual queries that are likely to cause problems in the near future. For each forecast, its validity (how likely it is to happen) and the time left until it's predicted to happen are shown. Clicking on View SQL will show the whole query text, while clicking on the object name will show the prediction's data points.