Infrastructure Planning, Sizing and Deployment

Infrastructure Planning

Node sizing

The following table shows the recommended node sizes for the Reporting pod/container depending on the Traffic Volume.
SrNo QPS on TML Cluster CPU for Reporting Node Memory for Reporting Node
1 1K - 3K 4 cores 15 GB

TML-Reporting stores everything on the disk attached to it, as it stores Prometheus data up to 1 year. So, you will need to assign a volume of large size to the TML-Reporting container/pod.

You can modify the storage capacity by modifying "tml_reporting_storage_size" in the manifest file.

TML-Reporting requires a dedicated node for its scheduling and execution. So, you should plan infrastructure for TML-Reporting in two ways if you choose to deploy the TML-Reporting later, once the TML-cluster is up and active.

This way would require you to have an extra node in the TML-cluster which would be used later to deploy TML-Reporting on it.

You can use the platform's capabilities to add an extra node to the existing running cluster to deploy TML-Reporting on it.

For TML-Reporting to be deployed with TML-cluster, then you should plan an extra node to the required nodes for it to be deployed.
Note: You need to label the nodes once the cluster is up as mentioned in Prerequisites for TML-Reporting deployment.