Use Case

12:30 PM - 1:00 PM, CEST , May 23

Pulsar Observability in High-Topic Cardinality Deployments for Telco

We have built our observability stack consisting of Grafana and Prometheus for our Cloud-Native Platform based on Apache Pulsar for the next-generation of Virtual CPE Onboarding and Secure Access Service Edge. Per the latter we have topic pairings that scale linearly per equipment and with each node generating approximately 1000 series. As we have passed the 50000 mark and aiming go past 1 Million of them, we have encountered major challenges in scaling of our monitoring stack and took it upon ourselves to extend the Grafana Agent Operator's Helm Charts along with a novel way of partitioning of labels and tenancy management to overcome this bottleneck. We would like to share our experience with the community.

Speaker

Habip Kenan Üsküdar

DevGitOps Engineer, Axon