Datadog introduces RUM Managed Archive, an extension of RUM without Limits. This feature lets you store user sessions cost-effectively, even if retention filters don’t select them. Teams can retrieve specific sessions later for support, incident investigation, or audits.
With Real User Monitoring (RUM), teams can see what users actually experience in web and mobile applications. Within the RUM without Limits model, Datadog already processes all traffic for metrics and SLOs. Retention filters determine which sessions are retained in full. Until now, sessions that did not meet these criteria were deleted.
This presents a dilemma. According to Datadog, indexing everything is usually too expensive. But selective retention risks missing a crucial session when a customer raises an issue weeks later. This is particularly problematic in regulated environments.
Managed Archive is designed to solve this. Instead of discarding unselected sessions, Datadog stores them in its own managed storage. According to the company, this is cheaper than indexing.
Storage configurable per application
The retention period can be configured per application: 1, 3, or 6 months. This allows teams to use different strategies for applications with varying audit or risk requirements. By comparison, indexed sessions are retained for 30 days by default in RUM without Limits.
Archived sessions are not fully indexed. However, they do contain a limited set of metadata, including a timestamp, session ID, user ID, and account ID. Using a separate view in the interface, users can search by these attributes to find an individual session or a group of sessions related to a specific incident or time frame.
When a session is retrieved, it is restored to RUM in its fully indexed form. It then enters a 30-day analysis window. A restored session offers the same investigative capabilities as a session retained via a retention filter, including access to events and attributes.
Support ticket weeks later
Datadog outlines an example from the financial sector. A customer places a trade via a mobile app, after which the confirmation screen freezes. A support ticket isn’t submitted until weeks later. The session fell outside all filters, so without an archive, the team would have nothing to investigate.
With Managed Archive, the team can retrieve the session using the user ID. What did they find? The user was running a version of the app with a UI update that delayed price refreshes, briefly showing an outdated price. The backend, however, executed the trade at the correct real-time price. The team later confirmed that a subsequent release resolved the issue.
In addition to support, Datadog mentions SREs who retrieve sessions from a specific time window for retrospective incident investigations and audits that require access to sessions beyond the standard retention period.
Tip: Datadog tackles AI and security complexity with new features