Gateway OpenTelemetry Collector recommendation
Proper sizing the number of Gateway OpenTelemetry required is crucial for optimal performance.
Telemetry data is collected from each node in the cluster using OpenTelemetry collectors and must be transmitted to the SaaS Observability solution. If outbound internet access is unavailable for all cluster nodes, specific Gateway nodes must be designated to centrally forward the telemetry data. These Gateway nodes are selected from the existing cluster nodes and must have outbound internet access. The OpenTelemetry collectors on these Gateway nodes will handle the transmission of the data.
High Availability enabled cluster
If you want to set up High Availability (HA) or have already set up HA on your cluster, Cloudera recommends a minimum of two Gateway nodes with
OpenTelemetry Collectors, irrespective of the number of nodes.
Scaling recommendation
The table below provides the recommendations for appropriate scaling.
| Number of Nodes | Count of Gateway OpenTelemetry Collectors |
|---|---|
| 100 | 2 |
| 500 | 2 |
| 1000 | 2 |
| 2000 | 3 |
| 4000 | 5 |
The formula to calculate the number of Gateway OpenTelemetry Collectors = 1 + ⌈(N/1000)⌉
where N is the number of nodes on which the OpenTelemetry Collector is present.
