# Are model calls, tool calls and the identity behind each logged and kept?

Facts read 16 Nov 2026. Page reviewed 16 Nov 2026.

Question 28 of the assessment asks: Are model calls, tool calls and the identity behind each logged and kept? When an agent acts, the log is the only record of what it was asked and what it did. It comes from Microsoft cloud security benchmark AI-6 and carries weight 2 in the ai workloads domain.

## Why it matters

When an agent acts, the log is the only record of what it was asked and what it did.

## Answers and what each is worth

| Answer | Value |
| --- | --- |
| Yes, logged and retained | 1 |
| Partly logged | 0.5 |
| No or not known | 0 |

## How to check it in each cloud

- **AWS:** Check Bedrock model invocation logging to S3 or CloudWatch Logs, and CloudTrail data events for agent and model APIs.
- **Azure:** Check diagnostic settings on Azure OpenAI and AI Foundry resources send request and audit logs to a Log Analytics workspace.
- **Google Cloud:** Check Vertex AI request-response logging and that Data Access audit logs are on for the Vertex AI API.

## The fix

Turn on invocation and tool-call logging for every AI service and keep it with your other audit logs. Effort: low.

## Where it comes from

- Microsoft cloud security benchmark AI-6: [Establish monitoring and detection](https://learn.microsoft.com/en-us/security/benchmark/azure/mcsb-v2-artificial-intelligence-security)
- NIST SP 800-53 Rev. 5 AU-2: [source](https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final)

## Related pages

- [Question 29: Guardrails and drift, preventive policies](/questions/preventive-policies/)
- [Question 30: Guardrails and drift, changes as code](/questions/changes-as-code/)
- [Question 31: Guardrails and drift, drift detection](/questions/drift-detection/)

## Sources

- Microsoft cloud security benchmark AI-6: https://learn.microsoft.com/en-us/security/benchmark/azure/mcsb-v2-artificial-intelligence-security (read 16 Nov 2026)
- NIST SP 800-53 Rev. 5 AU-2: https://csrc.nist.gov/pubs/sp/800/53/r5/upd1/final (read 16 Nov 2026)

Last reviewed 16 Nov 2026. Canonical: https://cloudsecurityriskassessment.com/questions/ai-activity-logging/
