CNCF-hosted Co-located Events North America 2024

Ensuring observability is critical, especially as operations expand into edge cloud environments. Effective observability data volume management is crucial to benefit from observability while controlling costs (storage, network, processing). Naive data reduction (e.g.,changing all metrics freq from 30s to 5min, sampling traces) can harm SLAs. With Observability Volume Manager(OVM), IBM and RedHat have co-developed an open-source intelligent framework with dynamic, fine-grained control of observability data volume. It works across modalities and integrates with open-source tools like Prometheus and OTel. OVM performs dynamic transformations at edge like filtering/aggregating data to reduce volume as well as adjust metric frequency/log levels for risk based zoom in/out and issue triaging. OVM identifies highly correlated/computed metrics and performs intelligent pruning and given an observability budget, identifies the subset of metrics and frequencies to minimize downstream task impact.