View this page in the latest version of Appian. Logging Share Share via LinkedIn Reddit Email Copy Link Print On This Page Purpose The Logging page in the Admin Console allows you to configure logging for integrations to facilitate troubleshooting. Access To access the Logging page, go to the Admin Console. Then, in the left navigation menu, under Integration, click Logging. Settings Trace Logging This section controls the integration_trace.csv, web_api_trace.csv, and appian_platform_api_trace.csv logs. When enabled, Appian will log tracing information about integrations, web APIs, and platform API requests. For example: if the request was successful or if it logged out. The Appian Platform API trace logging checkbox is only visible when the feature is enabled for your environment. See the integration, web API, and platform API sections on the Logging page for the fields that are captured. Tip: The only potentially sensitive data that is captured in the trace log is the username of the user who called the integration. HTTP Request/Response Logging This section controls the integration_req_resp_activity.log and web_api_req_resp_activity.log logs. When enabled, logs the raw HTTP request and response messages for integrations and web APIs that have logging enabled. Developers can turn on logging for any integration object that uses an HTTP or OpenAPI connected system, as well as for any web API object. See the Logging page for more information on the HTTP request/response logs. Note: The HTTP request/response logs may contain sensitive data or credentials. Before enabling HTTP request/response logs, it's important to understand the integration logging guidelines and web API logging guidelines. AI Input/Output Logging This section controls the AI input/output log (perflogs/ai/ai_input_output_activity.log). This log helps you monitor and audit AI usage within your Appian environment. Scroll or drag to pan · ⌘+/⌘− to zoom · Esc to close When the Enable input/output logging box is checked, the audit log captures additional details, including the inputs provided to the AI model and its response. It also records the specific generative AI Model ID used for the execution. This level of transparency empowers you to maintain control over AI usage and ensures compliance with evolving security and privacy standards. The following features are not captured in the AI input/output log: Process model node suggestions Inline testing in the AI skill designer Note: The input/output logs may contain sensitive data or credentials. Before enabling input/ouput logging, it's important to understand the AI input/output logging safeguards below. Logging safeguards There are a number of safeguards in place to ensure that AI input/output logging doesn't store sensitive data when used correctly: Administrator-only access: The APPIAN_HOME/logs/perflogs/ai_input_output_activity.log log is only accessible by system administrators. This prevents any privilege escalations for designers who shouldn't be able to see certain integrations. Log rollover policy: The log files roll over based on time and size. No AI input/output log will ever last more than 7 days. The cumulative size of all rollover over AI input/output logs will never exceed 100 MB. To learn more about the log rollover policy, see the Logging page. Avoid enabling input/output logging if: Inputs or outputs include sensitive information (e.g., PII or credentials) without proper safeguards. The data is not necessary for auditing or monitoring purposes. Tip: See the Logging page for information about the structure of AI Input/Output logs. Feedback Was this page helpful? SHARE FEEDBACK Loading...