The interval menu controls the time interval to display in the graphs. The precision of the graphs is determined by the interval. Some features are not available when the interval is over 12 hours.
You can also control the interval by clicking and dragging on a graph to highlight a section of the graph to zoom in on.
The start time can be set to control when the interval being shown starts at. If the start time is left blank, then it will be calculated as the most recent interval (i.e. if the interval is 30 minutes, the graphs will show the last 30 minutes).
If the live checkbox is checked, then the server will be polled every few seconds for update to the graphs. Live updates are only available if the interval is an hour or less.
The host menu can be used to filter the graphs for just specific hosts. It can be used to narrow down if any anonmalies are isolated to a specific host or not.
The action menu can be used to filter the graphs for just specific actions. It can be used to narrow down if any anonmalies are isolated to a specific host or not. Only the top 100 actions are shown.
The app menu can be used to switch between different apps that are being monitored.
The app menu will not appear if there is only one app being monitored.
The environment menu can be used to switch between different environments that are being monitored.
The environment menu will not appear if there is only one environment being monitored.
You can use the refresh button to refresh the graphs.
You can use the pagination (arrow) buttons to navigate forward and backward from the current start time.
The request time graph shows a detailed breakdown of how time was spent on average processing each request. You can use the request metric menu to switch which metric you want to see.
If you choose to see the average time, then the graph is broken down by component so you can see where the application is spending most of the request time.
You can also choose to see the p50, p90, or p99 metrics for response time. Due to the way that data is stored, these are not true p50, p90, and p99 values. Rather they are an average of those values over each process reporting data and then further averaged over the time frame. This should be accurate enough for most purposes. However, you could get different values if you gather these metrics from another source
The throughput graph shows the number of requests per minute processed by the application.
Errors are show in three different displays. You can see the error rate which is the number of errors encountered per request as well as the total number of errors. Both of these are show graphed over time.
You can also click the Show Error Details button to show a table with all the errors that were encountered over the time range along with their stack traces.
Clicking the Show Action Details button will show an additional graph that breaks down the top 25 actions that are taking up the most wall clock time to execute. This information can be used to detect hot spots in the code that can provide a benefit from optimization.
You can download the current data being displayed as a CSV file by clicking the Download Data button.