> For the complete documentation index, see [llms.txt](https://knowledgebase.autorabit.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://knowledgebase.autorabit.com/product-guides/guard/getting-started-with-guard/data-retention.md). # Data Retention {% hint style="info" %} These are the default retention windows for Guard. Contact [AutoRABIT Support](https://support.autorabit.com) if you need to discuss custom retention settings for your organization. {% endhint %} ### Compliance and audit records These records have the longest retention windows to support multi-year compliance reviews and audit trails.
FeatureDefaultWhat Guard keeps, and what happens when it expires
Change Log5 yearsGuard's record of every metadata change detected in your Salesforce org: what changed, when, by whom, any notifications Guard sent, and any baseline policy deviations it triggered. After 5 years, change entries are removed from the Change Log screens, exports and deviation history.
Guard audit trail5 yearsA versioned history of every edit made inside Guard itself: user accounts, connected hosts, subscriptions, saved queries, baseline policies, notification settings and AI usage limits. After 5 years, even the record that something was changed or deleted is gone permanently.
Code scan gate verdicts5 yearsThe pass/fail decision for every pull request Guard evaluated. These are the compliance receipts that prove a merge was gated. Verdicts are kept for 5 years, well beyond the scan results themselves (6 months), so you can still answer "was this merge gated?" long after the scan has aged out.
### Billing records | Feature | Default | What Guard keeps, and what happens when it expires | | --------------- | --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | AI credit usage | 13 months | The day-by-day record of AI credits consumed per feature. This is the data used for billing reconciliation. 13 months covers a full billing year plus a one-month buffer. | ### Feature history These records power Guard's timeline views and trend analysis. After one year, historical states are not recoverable. | Feature | Default | What Guard keeps, and what happens when it expires | | ---------------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | User Access Monitoring | 1 year | Periodic snapshots of every Salesforce user in your org: their profile, permission sets, licenses and attributes at each point in time. After 1 year, access history cannot be recovered. | | Risk Assessment | 1 year | Results of Guard's recurring security posture scans: overall scores, identified risks and their state at each scan. After 1 year, trend charts and historical assessments are truncated. Your current assessment is always regenerated by the next scan. | | API Security | 1 year | Periodic inventories of connected apps and API access posture in your org. After 1 year, connected-app history is removed. The current inventory is refreshed by every new snapshot. | | Data Classification | 1 year | Guard's AI-driven analyses of which objects and fields in your org hold sensitive data. Only superseded analyses age out — the most recent analysis per org is always kept, regardless of age. | | Drift Policy evaluations | 1 year | The outcome of every drift policy evaluation: whether your org drifted from its expected state, which rules triggered, and what the values were. After 1 year, evaluation history is truncated. The policy itself and its configuration are kept indefinitely. | | Transaction Security Policy events | 1 year | Events ingested when a Salesforce Transaction Security Policy fires in your org (for example a blocked login or data export attempt). The event feed reaches back 1 year; your TSP policies are not affected. | | Usage statistics | 1 year | Operational counters showing Salesforce API and AI/LLM calls Guard made on your behalf, per feature. Used for consumption dashboards. Billing data is handled separately by the AI credit ledger above. | | Digital Experience scans | 1 year | Results of Guard scanning your Experience Cloud sites for public exposure. After 1 year, scan history is removed. Each new scan re-establishes the current exposure picture. | ### Operational data | Feature | Default | What Guard keeps, and what happens when it expires | | ---------------------- | -------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | CLI error logs | 3 months | Error telemetry uploaded when a Guard CLI installation encounters a bug. Used by support and engineering to diagnose issues. | | Code scan debug traces | 3 months | Step-by-step scan traces the CLI can upload for support to diagnose a specific scan run. Deliberately shorter than scan results — a trace is only useful for recent runs. | | AI response cache | 2 months | Stored AI answers reused to avoid resending identical prompts. An expired cache entry has no functional impact — Guard recomputes the answer on next use. | | Expired API keys | 30 days | Keys that have passed their expiry date. The record is kept as a grace period during which an admin can still rotate the key. After 30 days, the key is deleted and a new key must be created and distributed to any pipelines using it. | | Revoked API keys | 30 days | Keys an admin has explicitly revoked. The record is kept for 30 days so the key list still shows what was revoked and when. Who revoked it remains visible via the audit trail for 5 years. | ### Temporary files Housekeeping cleanups for files that can be regenerated. | Feature | Default | Notes | | ---------------------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------- | | Generated export files | 5 days | Reports and exports (CSV, Excel, PDF) held on disk until downloaded. Download links older than 5 days stop working; simply re-run the export. | --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://knowledgebase.autorabit.com/product-guides/guard/getting-started-with-guard/data-retention.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.