When rolling out new features, problems can arise that did not occur in test environments. Datadog cannot put an end to that problem, but it can make it more manageable. To provide development and production environments with unambiguous observability, Datadog is now introducing Feature Flags.
Many feature flagging tools still operate in isolation from observability tooling, Datadog notes. This forces engineers to monitor changes across multiple disconnected systems to understand the full impact of issues. This slows down the development cycle and increases the likelihood of missed issues. Datadog wants to address this problem. Developers can now add feature flags to the observability platform used by IT teams. This allows health metrics to be correlated and the gradual rollout of new releases to be automated.
Datadog Feature Flags connects each flag directly to existing telemetry data. All metrics, logs, traces, and business KPIs are therefore provided with the flag. Engineers no longer need to set up integrations or write custom scripts for this functionality, because each rollout is automatically tracked against the health indicators that are already being monitored. When rolling out a new checkout experience, IT professionals can see in real time how this affects response times, error messages, and conversion metrics.
Building on OpenFeature
The solution builds on the OpenFeature SDK and enables teams to quickly create flags and variants with any data type: Boolean, string, number, or JSON. Within a newly created flag, teams can add the flag to a specific codebase, define targeting rules for rollout in specific environments, add variants, and view real-time metrics.
A common problem for IT teams is that they have to switch between dashboards to measure application health or user behavior. Whether a change within that is parallel to the rollout of a new release or feature is usually unclear. To speed up the triage of problems, Datadog now makes it possible to locate and disable the relevant flag. That step is also made clear in the logging.
Automation and rollback
Datadog’s ultimate goal is to automate the rollout of new releases and features. This means that regardless of the presence of problems, IT teams retain an overview. This eliminates the guesswork surrounding bugs and other issues.
Datadog notes that the extra peace of mind in the rollout cycle increases the opportunity for experimentation. It becomes clear more quickly whether a new release is operating as intended, allowing development teams to move faster toward new innovations.
Read also: Datadog acquires Metaplane, a data problem-solver