I've been knee-deep in evaluating Chronosphere against other platforms like Datadog and Grafana Cloud for the last quarter, and while the cost predictability is a major draw, the configuration complexity is a real barrier.
I'm not talking about the UI—it's the underlying concepts and sheer number of moving parts. To get something "simple" like a custom aggregation rule or a cost-controlled metric stream, you're juggling:
- **Control plane** vs **data plane** resources (and their separate APIs)
- **Bucketization rules** that feel like you're writing a mini-ETL job
- **Mapping rules** to route metrics, which have a different syntax and logic than the bucket rules
- **Resource policies** to then govern those mapped streams
It feels like you need a dedicated "Chronosphere engineer" on staff just to model your data correctly before you even start observing anything. With other platforms, you ingest and then you *maybe* tweak retention or sampling. Here, you have to pre-define your entire ingestion taxonomy.
Has anyone built a streamlined internal workflow or a set of templates to make this more manageable? I'm specifically trying to avoid a scenario where we misconfigure a mapping and blow through our contracted usage budget because a high-cardinality metric wasn't properly filtered.
I've seen the docs and the examples, but they all seem to assume you have a small, static set of well-behaved metrics. Our environment is dynamic, with services spinning up/down constantly. How are other teams handling this without dedicating a full-time role to Chronosphere config management?