Documentation Index
Fetch the complete documentation index at: https://docs.odigos.io/llms.txt
Use this file to discover all available pages before exploring further.
Operational and design considerations to keep in mind when running sampling in production.
Gateway load, OOM, and scaling
When the gateway accepts a trace for tail evaluation, it allocates memory to hold spans until the aggregation window closes. If the gateway isn’t sized correctly, bursts of traffic can cause OOM restarts. Symptoms include dropped spans or collector crashes that aren’t caused by your sampling rules.
For tail sampling, prefer scaling up (bigger collector pods) over scaling out (more replicas). All spans of a single trace must land on the same pod for aggregation, so fewer, larger pods simplify routing and load balancing. Validate sizing against your own load—there’s no one-size-fits-all.
When any sampler enables tail mode
Any sampling rule that requires tail evaluation activates tail mode for the entire cluster, even rules that look inactive or minimal—for example a placeholder service name or a broad / route. Once tail mode is on, the gateway buffers each trace for the aggregation window, which increases memory use, export latency, and the risk of dropped spans. If you don’t need tail behavior, either remove any unused or test sampling rules from the cluster, or disable tail sampling explicitly in Odigos configuration (Helm values, or the UI Settings page). 
