Clari incident

RevDB Loading Delays

Major Resolved View vendor source →

Clari experienced a major incident on January 5, 2026 affecting Revenue Database (RevDB), lasting 5h 43m. The incident has been resolved; the full update timeline is below.

Started
Jan 05, 2026, 04:32 PM UTC
Resolved
Jan 05, 2026, 10:15 PM UTC
Duration
5h 43m
Detected by Pingoru
Jan 05, 2026, 04:32 PM UTC

Affected components

Revenue Database (RevDB)

Update timeline

  1. investigating Jan 05, 2026, 04:32 PM UTC

    We are investigating delays in loading CRM and Activity data for modules that leverage RevDB.

  2. investigating Jan 05, 2026, 05:31 PM UTC

    The investigation into RevDB loading delays is ongoing. Customers may see delays loading new CRM and Activity data.

  3. investigating Jan 05, 2026, 06:22 PM UTC

    Investigation and mitigation continues. Customers may still see delays loading new CRM and Activity data.

  4. monitoring Jan 05, 2026, 08:05 PM UTC

    The root cause of the loading delays has been identified and fixed. Data loading for customers is catching back up and we continue to monitor the system.

  5. resolved Jan 05, 2026, 10:15 PM UTC

    Data loading delays have been resolved.

  6. postmortem Jan 07, 2026, 10:08 PM UTC

    On 2026-01-05 at approximately 13:50 UTC, Clari experienced an interruption in the processing of CRM and Activity data into RevDB. Due to database locking issues across multiple clusters, live CRM and Activity updates stalled for some customers, resulting in data freshness delays of roughly 2 to 5 hours. The impact was limited to delayed data updates; the application experience was unaffected, and no customer data was lost. Clari engineers identified a buildup of long-running, idle-in-transaction database sessions that were causing lock contention. The problematic sessions were terminated, unblocking the affected clusters and allowing CRM and Activity ingestion to resume. To reduce the likelihood of this issue recurring, Clari has implemented changes to prevent idle sessions from accumulating, and we are strengthening monitoring and alerting for database lock behavior to detect and quickly mitigate similar conditions.