Few good practices for running in scale: 1. If you are on k8s, start with k8s helm charts - easier to scale with them 2. Create a separate cluster for running SIgNoz, in this way the impact of something going wrong in your app layer is isolated 3. start slow, start with few services first see how much data it is ingesting and based on that set up the disk you want to allocate 4. Set retention periods for metrics, traces and logs - and make sure the storage you allocate is able to handle that