When an application on Acquia Cloud Platform slows or returns an error, users can delegate the investigation to Acquia AI. The agent analyzes logs and metrics for the specified environment and time window, correlates evidence, and reports whether the cause resides in the application or the platform. Where a fix applies, the agent recommends next steps for code owners to apply.

Acquia AI responds to user requests and does not monitor environments independently.

Prerequisites

Ensure that the setup meets the following requirements:

• The organization is entitled to Acquia AI, and an administrator account granted access.

• The target application runs on Acquia Cloud Platform. Applications hosted outside Cloud Platform receive no coverage.

When an application on Acquia Cloud Platform slows or returns an error, users can delegate the investigation to Acquia AI. The agent analyzes logs and metrics for the specified environment and time window, correlates evidence, and reports whether the cause resides in the application or the platform. Where a fix applies, the agent recommends next steps for code owners to apply.

Acquia AI responds to user requests and does not monitor environments independently.

Prerequisites

Ensure that the setup meets the following requirements:

• The organization is entitled to Acquia AI, and an administrator account granted access.

• The target application runs on Acquia Cloud Platform. Applications hosted outside Cloud Platform receive no coverage.

• An administrator assigned the target application to the project agent.

• The subscription holds available AI credits. For details, visit AI credits and usage.

Existing Cloud Platform sites require no new instrumentation. Cloud Platform trial accounts maintain the same diagnostic functions as paid subscriptions.

Delegate a diagnosis

To delegate a diagnostic task:

1. Open Acquia AI and set the mode toggle to Work mode.
   Diagnosis functions operate in Work mode only.

2. Name the application and environment, describe the symptom, and specify the target time window. For example: On the production environment of the marketing application, the error rate rose sharply after 09:00 today. What does the evidence show?

3. Submit the request.

4. Wait while the agent retrieves logs.

5. Review the returned findings and recommended next steps.

Specify one environment per request. The agent analyzes the named environment rather than sets of environments simultaneously.

Log processing expectations

Log retrieval from Cloud Platform requires multiple steps: the agent requests the log, the platform generates it and returns a link, the agent retrieves the file, and analysis begins. Larger logs extend retrieval times, and diagnoses typically require multiple logs. Consequently, diagnostic tasks require more time than queries in Advisor mode.

Two practices optimize analysis time:

• Specify narrow time windows: Select time windows of a few hours around symptoms to yield faster results than multi-day windows.

• Specify environments precisely: Precise environment selection prevents the agent from retrieving unnecessary data.

Analyzed data sources

For specified environments and time windows, the agent analyzes:

• Apache access and error logs

• Cloud hook logs

• Cron logs

• Drupal request and watchdog logs

• PHP FastCGI Process Manager (FPM) access and error logs

• PHP error logs

• Slow query logs

Diagnostic output

The agent returns the following details:

• Source layer attribution: application or platform.

• Correlated evidence compiled in a central location.

• Scope parameters specifying evaluated and excluded data to establish finding context.

• Recommended next steps where fixes apply.

• Escalation guidance to Acquia Support when appropriate.

Raw log content does not appear in responses. The agent reports identified patterns and evidence rather than individual log lines. To examine raw logs, retrieve them directly through the Acquia Cloud Platform interface.

Diagnostic boundaries

The agent cannot perform the following actions:

• Take corrective actions based on findings.

• Modify application code.

• Modify infrastructure configurations, such as cache clears, log forwarding changes, or certificate replacements.

• Submit support cases.

• Resolve outages or critical issues.

For additional details, refer to Governance and control management.

Interpret diagnostic evidence

Performance issues involve broad symptoms, and attribution errors waste operational time. Review the evaluated log scope and weigh recommendations against known application context before applying changes.

Steps after diagnosis

Follow these procedures based on diagnostic findings:

• Platform findings: Submit a case to Acquia Support and include the diagnostic findings.

• Application findings: Forward the findings and recommended steps to the code owner.

• Page-specific recommendations: Run a governance check on affected pages if recommendations target specific pages rather than the environment. For additional details, visit Run a governance check and apply a fix.