Is Permutive Down?

This incident has been fully resolved. All delayed data from 19 September onwards has been loaded, and reporting in Insights, Live Audience Stats and Plans is now up to date. No data was lost during this period. The delay was caused by a change to a cloud service we use to process reporting data. We have since changed how our pipelines use that service so that this type of change can no longer disrupt processing. Thank you for your patience. --- [2026-09-22T12:52:08.076Z] (monitoring) Recovery is nearly complete. The majority of the delayed data from 19 September onwards is now visible in Insights, Live Audience Stats and Plans, and the final portion is being loaded. Some customers may still see a small gap for the most recent days until this finishes. No data has been lost. We will post a final update once we have confirmed that reporting is fully up to date for all customers. --- [2026-09-21T15:18:31.147Z] (monitoring) The underlying issue has been resolved, and data ingestion has resumed. We are now loading the delayed data from 19 September onwards, and it will progressively appear in Insights, Live Audience Stats and Plans as it is processed. No data has been lost. We will continue monitoring and post a further update once reporting is fully up to date. --- [2026-09-21T09:59:09.559Z] (identified) Since 19 September 2026, the data shown in Insights (Audience & Campaign), Live Audience Stats and Plans has been delayed. Reporting in these products currently stops at 19 September and does not yet include the most recent days. The delay is caused by an issue with a cloud service we rely on to process reporting data. Our engineering team is working with the provider to resolve it. No data has been lost. Your events continue to be collected as normal, and the missing days will appear automatically once processing resumes. Audience targeting, activation and all other Permutive features are unaffected. We will post a further update as soon as possible. We apologise for the inconvenience.

Between 21:14 and 22:13 UTC on 3 September, the CCS (Custom Cohort Segmentation) API experienced degraded performance. During this period the large majority of segmentation requests failed with HTTP 500 errors after a delay of several seconds. Events sent in the affected requests were still collected. However, users were not segmented and no cohorts were returned from those requests. This only affected segmentation done via direct calls to the CCS API. Requests outside this window were unaffected. The issue is now resolved and the API has returned to normal. This was caused by a failure in an upstream service which impacted the CCS API. Improvements to both the upstream service and the CCS API have been identified and are being worked on which will prevent a reoccurrence of this failure and its impact.

Between 14:10 and 14:30 UTC, the Permutive API experienced an outage. During this period we were unable to receive events and identities. User targeting was unaffected. However, events sent during this window were not collected, and will therefore be missing from two areas: Routing Insights Data collected outside this window is unaffected. The issue is now resolved and event collection has returned to normal. --- [2026-09-01T14:35:22.429Z] (monitoring) A fix has been implemented and we're continuing to monitor. --- [2026-09-01T14:27:16.000Z] (investigating) We are investigating an outage of Permutive API.