Hey everyone. So I'm pretty new to using Traceloop and just got my first real bill. It was... a lot higher than I expected 😅
I’m still learning cloud cost management. What are the main things that drive up costs in Traceloop? Is it the number of traces ingested, the data retention, or something else? And what's the best way to audit what happened? I'm on AWS but also curious about GCP/Azure.
Still learning
The bill shock is a rite of passage in observability platforms. For Traceloop specifically, the primary cost driver is almost always the volume of trace data ingested. Retention period matters, but ingestion is the real variable cost.
You need to look at your span volume and any sampling configuration (or lack thereof). A common trap is sending debug/trace logs as spans or instrumenting at too verbose a level (e.g., every database fetch as its own span). Check what your instrumentation libraries are emitting.
On AWS, immediately pull your CloudWatch metrics for the Traceloop namespace (if they provide them) or use their API to get usage stats for the billing period. Segment it by environment (prod vs. dev) and service. You'll likely find one service or one environment is responsible for 80% of the volume. Implementing head-based sampling in your tracing SDK before data leaves your hosts is the standard fix.
infrastructure is code