AI Trust Layer Audit Log Degradation in the US Region

Update timeline

1. investigating Aug 05, 2026, 11:28 PM UTC

   We are investigating reports of degraded performance impacting visibility of recent audit log records for AI Trust Layer in the US region. Impact: Users may experience missing or delayed recent records on the AI Trust Layer Audit Log page. Audit log ingestion remains healthy. Our teams are working to identify the cause and will share more details as the investigation progresses.

2. identified Aug 05, 2026, 11:40 PM UTC

   We have identified the cause of degraded performance impacting visibility of recent audit log records for AI Trust Layer in the US region. Impact: Users may experience missing or delayed recent records on the AI Trust Layer Audit Log page. Audit log ingestion remains healthy. Our teams are redeploying the affected service and will continue monitoring recovery. We will share another update as soon as service behavior has stabilized.

3. identified Aug 06, 2026, 12:40 AM UTC

   We have applied a fix and are seeing the affected aggregation jobs run successfully again. Impact: Users may continue to experience delayed visibility of recent records on the AI Trust Layer Audit Log page while processing catches up. Audit log ingestion remains healthy. Our teams are monitoring the backfill process and will share another update once audit log visibility has fully recovered.

4. identified Aug 06, 2026, 01:51 AM UTC

   The fix remains in place and the affected aggregation jobs are continuing to run successfully. Impact: Users may still experience delayed visibility of recent records on the AI Trust Layer Audit Log page while the backfill continues. Audit log ingestion remains healthy. Our teams are continuing to monitor the backfill progress and will share another update once audit log visibility has fully recovered.

5. monitoring Aug 06, 2026, 03:48 AM UTC

   The issue impacting visibility of recent audit log records for AI Trust Layer in the US has been resolved. The affected aggregation jobs are running successfully, and backfill processing has returned to normal. Audit log ingestion remained healthy throughout the incident. We will continue to monitor service health. Thank you for your patience.

6. resolved Aug 06, 2026, 05:20 AM UTC

   All functionality has been restored and the issue is now resolved. Impact: Users should no longer experience issues.

7. postmortem Aug 18, 2026, 09:26 PM UTC

   ## Customer impact Between August 3, 2026 at 3:27 pm UTC and August 6, 2026 at 3:49 am UTC, customers in the US region experienced missing or delayed recent records on the AI Trust Layer Audit Log page. The customer-facing impact lasted approximately 60 hours and 22 minutes. Audit log ingestion remained healthy throughout the incident — records continued to be received and stored. The issue affected only the visibility of recent audit log records while the background processing that prepares records for display was unavailable or recovering. ## Root cause The root cause was a failure in the background processing that prepares recent AI Trust Layer audit log records for display. The processing was failing due to database query timeouts, which prevented newly ingested records from appearing on the Audit Log page. A service restart resolved the immediate failure, after which recovery processing was needed to catch up on the backlog of unprocessed records. ## Detection The audit log view first stopped updating at 3:27 pm UTC on August 3, 2026. The issue was reported and an incident was created on August 5, 2026 at 11:12 pm UTC — approximately 55 hours and 45 minutes after impact began. The incident was declared customer-impacting at 11:13 pm UTC. Investigation confirmed the issue affected visibility of recent AI Trust Layer audit log records in the US region. ## Response At 11:28 pm UTC on August 5, we published an initial status update noting degraded visibility of recent AI Trust Layer audit log records. At 11:40 pm UTC, we confirmed the cause and communicated that the affected service was being restarted. The restart had begun by 11:31 pm UTC. At 12:25 am UTC on August 6, the service restart completed and the background processing was confirmed to be running successfully. At 12:40 am UTC, we updated our status page to confirm the fix was in place and that recovery processing had started, while noting that some customers might still see delayed records until processing caught up. At 3:49 am UTC, recovery processing had returned to normal and recent audit log visibility was fully restored. We continued monitoring and marked the incident resolved at 5:20 am UTC after confirming no further issues. ## Follow up 1. We are evaluating follow-up actions to improve detection speed and prevent recurrence, informed by the processing failure behavior and the recovery observed during this incident. 2. We are reviewing our monitoring and alerting coverage for this service to reduce the time between impact onset and incident detection.