Documentation forSolarWinds Observability SaaS

Service transactions

A transaction identifies a particular type of work within an application. For web applications, this typically means a particular controller/action, a route, or an individual endpoint for an API. When you install the APM libraries on a supported framework, transactions are automatically identified. For frameworks that are not supported or if the automatically derived transaction name does not suit your needs, a custom transaction name can be set through configuration or instrumentation SDK.

Customizing the transaction name

All APM libraries support setting a custom transaction name through the instrumentation SDK. Alternatively, configurable transaction naming is available for some APM libraries. See the section below for supported languages.

The transaction naming precedence in descending order is:

  1. Custom name set with the SDK

  2. Configured name

  3. Automatic naming

The following flowchart illustrates the process.

Set the transaction name via SDK

For information about using the SDK to set the transaction name, see the following APM library instrumentation SDK sections:

Configurable transaction naming

Configurable transaction naming allows specifying a static name via environment variable, or using span attributes to form the transaction name for a request. For more information, see the following APM library configuration section:

View detailed information about a transaction  

From the Entity Explorer or the Application Performance Monitoring area overview, click the name of a service to open the service details view, then click the Transactions tab.

Clicking a metric for a selected transaction in the table opens the Inspector Panel, which displays the request rate, response time of the transaction, and error ratio. Click View Details in the Inspector Panel or the name of the transaction in the Transaction tab to open the Transaction details view.

Overview

The Overview tab provides key performance metrics about the selected transaction, and includes the following capabilities and widgets:

Analyze Transactions with APM AI Assist

Click the Analyze button to use APM AI Assist for Transactions. Built on SW1, AI Assist for Transactions generates a Transaction Performance Assessment report that analyzes the performance of a selected transaction. The report evaluates latency, errors, throughput, and related telemetry to identify performance bottlenecks, summarize key findings, and provide prioritized recommended actions to help you optimize transaction performance and reliability.

When the analysis is completed, AI Assist returns the following information in each assessment: 

  1. Executive Summary

    Overall health rating, one-sentence primary cause, and up to 3 Top Operational Risks tagged with Critical, High, Medium, and Healthy severity icons (). The Executive Summary is a high-level overview capped at 120 words.

  2. Health Signals Overview

    The signals table displays Latency (Avg/Median/P95/P99), Error Rate, Throughput, Traffic Trend, Anomalies, Active Alerts, and Dependency Bottleneck (only rows with real data), followed by a dependency summary listing telemetry-backed dependencies with health status.

  3. Key Findings

    Up to 5 findings grouped under Performance, Reliability, Throughput, Dependency Observations, Historical Anomalies, Active Alerts, or Other. Each finding includes:

    • What is happening

    • Evidence

    • Impact

    • Optional Dependency Impact

    • Likely/ Possible Explanation

    • Recommended Next Step

    • Priority and Confidence ratings

  4. Recommended Actions

    Three time-boxed recommended action tables are provided (Immediate 0–24 hours, Short term 1–7 days, Preventive 1–4 weeks), each with Priority, Action, and Expected Outcome columns, tied directly to specific findings.

  5. Final Assessment

    One-line Assessment, Biggest Risk, and Best Next Diagnostic Step, plus a Scope of Analysis subsection covering the time window, data volume, and dataset limitations.

These results are presented directly in the SW1 panel, making it easy for you to continue the conversation with the AI Agent about the returned analysis to get even deeper conclusions.

Average Trace Response Time Breakdown

The Average Trace Response Time Breakdown widget shows how much of the total response time was spent in each traced operation in the transaction performed by the service. Hover over the chart to see a quick summary for the aggregated response time within the selected interval.

Status Codes

Status Codes widget shows the HTTP status codes issued in response to the selected requests. Status codes are classified by the first number in the code:

  • 1xx: Informational
  • 2xx: Successful
  • 3xx: Redirection
  • 4xx: Client error
  • 5xx: Server error

Request Rate

The Request Rate widget shows the average request rate for the selected time period and the percentage of increase or decrease from the previous time period. The line graph shows the request rate values recorded over the selected time period.

The request rate is the number of requests completed per second.

Response Time

The Response Time widget shows the amount of time between the request being made to the result being returned. The line graph shows average response time and response time distribution in predefined percentiles (50th, 95th, and 99th).

The response time is the time it takes to process an inbound request in milliseconds.

Error Ratio

The Error Ratio widget shows the average error ratio for the selected time period and the percentage of increase or decrease from the previous time period. The line graph shows the error ratio values recorded over the selected time period.

The error ratio is the percentage of requests that result in an error. It includes all errors, not only unhandled errors.

Traces

On the Traces tab, the Traces Explorer opens to display a high-level trace overview related to the selected transaction, with links to details about individual traces. See Traces Explorer for more information about this area and its features.