Trace details page functions
The Trace details page displays a table with the timing data for the segments in a transaction, plus additional features and functions:
|Trace details page||Description|
|Primary trace data, functions||
The top of each transaction trace UI page shows several primary pieces of data, including the trace time, response time, and, if available, CPU info. Also, sometimes the UI may indicate a partial trace.
Available functions include:
The duration of each segment and subsegment is shown as both the number of milliseconds and a percentage of the total transaction time. Especially slow segments are color-coded.
To expand segments, use either of the following:
The Drilldown column may contain icons that link to additional transaction data, if available.
The Timestamp column displays the start time of a segment or subsegment, relative to the start of the transaction. For example, a value of
In the Drilldown column, icons indicate that additional data is available:
|Database queries||The database [database icon] icon links to the raw database query for that query segment. Depending on your transaction trace settings, this query can either show as raw data or obfuscated data. If the total time for the segment exceeds the query plan threshold in your settings, the query analysis also will appear, along with a stack trace showing the exact location of the call in your code.|
|Stack traces||The magnifying glass [magnifying glass icon] links to stack trace details.|
|Aggregate transaction details||This icon indicates that cross application tracing details are not available. However, you can select this icon to view aggregated transaction information across multiple invocations of the transaction.|
Color coding of segment duration data
The Duration column of the Trace details table is color-coded:
Red: More than 25% of the total transaction time was spent in this segment.
Yellow: More than 5% of the total transaction time was spent in this segment.
Color-coded durations do not necessarily indicate a problem. Whether the timing is actually a problem depends on what you consider normal and acceptable for your application.
Application code in traces
If a segment is listed as Application code in the table, it indicates that that section of code was not instrumented. You can use custom instrumentation to get more detail on that part of the code. Segments will often be listed as Application code for partial traces.
For performance reasons, New Relic APM only captures the first 2,000 segments of a transaction trace (900 for Node.js). If a transaction trace exceeds that number of segments, the agent truncates the transaction trace and displays a message at the top of the Details page:
This is a partial trace.
The New Relic APM agents have differing rules for when to truncate segments.
|APM agent||Truncate segment rules|
Truncates based on segment priority:
For more information, see the C SDK's documentation about
The segment duration must be greater than the
|Java, .NET, Ruby|
The Node.js agent only captures the first 900 segments of a transaction trace. For more information, see Node.js transaction tracer configuration.
Truncates based on segment duration; the slowest 2,000 segments are captured for the trace. For more information, see PHP transaction tracer configuration.
Truncates based on segment duration; the slowest 2,000 segments are captured for the trace. For more information, see Python transaction tracer configuration.
Segment grouping rules
Transaction traces that have a lot of segments can be hard to interpret. New Relic helps you by automatically grouping fast and/or repeated method calls into a single segment grouping.
For example, you might see a segment titled
7 fast method calls. This means that, because of their fast completion, New Relic assumed these calls were unlikely to be a problem and grouped them together. You can expand that grouping to see the individual segments.
Rules underlying segment grouping include:
- Fast calls: When New Relic detects four or more consecutive segments with a duration of 7.5 ms or less, these are grouped into a single collapsed header.
- Identical calls: When New Relic detects four or more consecutive identical calls, this usually indicates an N+1 problem. These segments are collapsed. Expanding the collapsed node displays only the first three calls, not all calls.