How Dubber modernized observability with Grafana Cloud and Adaptive Metrics

Company: Dubber
Industry: Telecommunications
Dubber is a global multi-cloud call capturing and conversation intelligence platform, processing over one million calls daily across 215+ partners worldwide. Managing observability across a fragmented mix of self-hosted tools — Grafana, Kibana, AWS OpenSearch, and AWS Managed Grafana — was consuming significant effort from a lean platform team. After migrating to Grafana Cloud, Dubber cut observability run costs by 25%, eliminated self-hosted maintenance overhead, and built a unified platform for metrics, logs, synthetics, and Kubernetes monitoring.
ChallengeDubber's observability stack had grown fragmented across four systems. Self-hosted infrastructure demanded constant patching, upgrades, and security management. Scaling compute to keep up with growing data volumes added further burden. Alert fatigue from over a million metrics series made separating signal from noise increasingly difficult. And with no single pane of glass, engineers switched between consoles during every incident.
SolutionAfter a proof-of-value exercise in Q4 2024, Dubber migrated to Grafana Cloud with support from Grafana Professional Services. The migration covered metrics and logs, browser-based synthetic monitoring, Kubernetes Monitoring across Azure and AWS, custom integrations (AWS CloudWatch for managed Kafka, Temporal for workflow monitoring), and AI usage dashboards tracking LLM token consumption by model and region. Adaptive Metrics reduced observability costs by eliminating noise across 1M+ metrics series.
"We also use this Kubernetes Monitoring for capacity planning. I did hear the word SaaS economics a lot. So like FinOps is something that I'm pretty close to. So we use this to also cost optimize some of our cloud spend based on the performance of our Kubernetes clusters and the compute."
– Sai Rampe, Head of Platform Engineering
ImpactDubber reduced observability run costs by 25% after migrating from four fragmented tools to Grafana Cloud.
1M+ metrics series optimized via Adaptive Metrics
Eliminated self-hosted maintenance (patching, upgrades, vulnerability management)
Single pane of glass across metrics, logs, synthetics, and Kubernetes
Synthetic monitoring probes catch P1 issues every 5 minutes before customers report them
CoreDNS dashboard identified and resolved a production DNS issue
Terraform-native management via Grafana Cloud IaC modules
Multi-stack architecture satisfying global data sovereignty and compliance requirements
"The beauty of Adaptive Metrics is it's all pre-built, it's all in the click of a button where all you have to do is just review it and then just click it."
— Sai Rampe, Head of Platform Engineering
Looking AheadDubber is completing their migration to Grafana IRM this quarter. Looking ahead, the team is evaluating tracing, profiles, frontend observability, and K6 performance testing — continuing to deepen Grafana Cloud as their strategic observability platform.
Speakers

Sai Haritha Rampe
Head of Platform Engineering, Dubber