Grafana Labs announced updates to its fully managed Grafana Cloud observability platform: The powerful new Adaptive Metrics feature, which enables teams to aggregate unused and partially used time series data to lower costs, is now available for broader public access. This feature leverages enhanced insights into metrics usage recently added to Grafana Cloud's Cardinality Management dashboards, which are now available in all Grafana Cloud tiers, both free and paid. Together these advancements, powered by the open source project Grafana Mimir, help organizations rapidly scale at cloud native pace while optimizing metric cardinality and controlling costs.

With the increasing usage of cloud native architectures and rapid adoption of Prometheus and Kubernetes, observability teams are faced with an explosion of metric data. Prometheus cardinality can often lead to surges in metric costs. Teams are looking for solutions that help them scale out metrics adoption within their company without massively growing their spend. To help solve this problem, Grafana Cloud now offers: 

Cardinality Management dashboards for identifying unused metrics

Grafana Cloud's Cardinality Management dashboards now include insights into the usage of high cardinality metrics, to help distinguish between metrics that are being used and metrics that are unused. The ability to identify high cardinality metrics that are unused in dashboards, queries, recording rules, and alerting rules results in actionable outcomes for SRE or centralized observability teams looking to confidently make data-driven decisions to reduce metrics spend without impacting observability. The Cardinality Management dashboards were first introduced late last year to Grafana Cloud Pro and Advanced customers, but now are generally available to all Grafana Cloud users, including those on the Grafana Cloud Free tier.

Adaptive Metrics for turning usage insights into cost savings

Grafana Cloud's Adaptive Metrics feature takes insights about usage from the Cardinality Management dashboards one step further: It gives users better control of spend on observability metrics by enabling aggregation of unused or partially used metrics. (With partially used metrics, only a subset of the metric's labels are used.) 

The Adaptive Metrics aggregation engine transforms these metrics into lower cardinality versions of themselves at ingestion. Unused or partially used labels are stripped from incoming metrics, reducing the total count of time series persisted - and thus the user's monthly bill.

Adaptive Metrics recommends aggregations based on an organization's historic usage patterns, and users can choose which aggregation rules to apply. Dashboards, alerts, and historic queries are guaranteed to continue to work as they did before aggregation, with no rewrites needed. If usage needs change, users can immediately revert back to the unaggregated version of a metric and get the extra detail they need going forward. 

Based on results reported by early users, Grafana Cloud Adaptive Metrics can eliminate an estimated 20-50% of an organization's time series with no perceived impact on their ability to observe their systems. 

"We carefully watch our metric and cost consumption, and in the past we manually evaluated every metric to identify what to drop, which was extremely time consuming and a tedious process. Grafana Cloud Adaptive Metrics simplified this process for us by generating recommendations curated for our environment, reducing the amount of time spent by half. I wish I had this feature sooner," said Lydia Clarke, DevOps Engineer at SailPoint.

Grafana Cloud Adaptive Metrics is now available in a public access program for all Grafana Cloud tiers. Sign up for access here. 

"While we've seen the value that Prometheus brings to organizations, we've also seen its popularity lead to rapid adoption and uncontrolled costs," said Tom Wilkie, CTO at Grafana Labs. "In fact, we even had this problem at Grafana Labs, running our own Prometheus monitoring for Grafana Cloud. One of our clusters had grown to over 100 million active series, and 50% of them were unused. We started thinking about how we could solve this problem, and Adaptive Metrics was the answer. We've reduced that cluster by 40%, and we're excited to share this powerful capability with our Grafana Cloud users."