AppDynamics categorizes the ten items that deviate the most from the baseline performance as node problems.
You can analyze node problems from the Node Problems viewer.
The Machine Agent must be installed on the machine hosting the node that you are troubleshooting.
To Access the Node Problem Viewer
To access the Node Problems viewer do one of the following:
- In the Node dashboard, from the Actions drop-down menu click Analyze Node Problems.
- In the left navigation panel of the Snapshot viewer, click NODE PROBLEMS.
Node Problem Viewer
The right panel of the Node Problem viewer displays the metrics that deviate the most from their baselines for the specified time range.
The left panel of the Node Problem viewer lets you filter the types of data reported as node problems.
If accessed from the Snapshot viewer, the Node Problem viewer displays node problems for the time range of the snapshot. If accessed from the Node dashboard, it uses the time range set in the Node dashboard. You can edit the time range in the Node Problem viewer and then apply the new time range. You can also define and save a custom time range.
The selection of metrics displayed depends on the configured filter options.
The following Node Problem viewer shows a transaction that took over 10 seconds (10069 ms) to execute.
It shows that the request count was 62 standard deviations from its normal baseline and that the DB connection pool was 1.6 standard deviations above its normal baseline. It maxed out at 30 concurrent connections.
The Hot Spots tab for this snapshot shows the details of the connection pool latency.
In this case, the Node Problems viewer reveals how excessive web requests saturated the DB connection pool.
Filter Options for Node Problem Analysis
You can filter the following options in the Filter Options panel of the Node Problem viewer:
Baseline to Use
Specify which baseline to use to define node problems:
- No baseline
- All data - Last 15 days
- Daily Trend - Last 30 days (default)
- Weekly Trend - Last 6 months
- Monthly Trend - Last year
For information about baselines, see Detect Anomalies Using Dynamic Baselines.
Data to Analyze
Select the type of data to analyze from the drop-down menu. AppDynamics analyzes the ten items that deviate the most from the baseline performance for the specified type.
Specify whether to display problems with values that are:
- higher than baseline
- lower than the baseline
- higher and lower than baseline
For example, if you are only interested in CPU that is too high, set the Analysis Type to Higher for Hardware Resources. On the other hand, if you want to monitor the load on your machine continuously because low CPU usage would also be a problem, set the Analysis Type to Higher and Lower.