# Tentative Goal Big Bang Core all running Highly Available consumes less than: * 10 vCPU * 20 GB Ram [P1BIGROCKS-2278](https://jira.il2.dso.mil/browse/P1BIGROCKS-2278) When using Big Bang and Zarf to coordinate and deploy a single application on an edge node some assumptions can be made to reduce the total compute footprint required for BigBang Core: | Big Bang Core Package | Deployed in Appliance Mode | Comments | |-------|----------|---| | Flux | Part | Just source and helm controllers | | Istio | Yes | Too heavy for reduced compute | | Jaeger | No | Not enough value to justify | | Tempo | Yes | integrated with grafana | | Kiali | No | Not enough value to justify | | Monitoring | Yes | Do we need Alertmanager as well? | | ECK | No | Too heavy for reduced compute | | Loki/Prom | Yes | need logging | | Gatekeeper/Kyverno | No | Static environment on edge, compliance will be validated in development/cloud | | Cluster Auditor | No | Static environment on edge, compliance will be validated in development/cloud | | Twistlock | Yes | Runtime security at least | This leaves a couple of gaps that need to be filled: ## Ingress ### Istio Istio/Istiod can be fairly heavyweight since istiod is expected to have [1CPU/1.5GB of resources](https://istio.io/latest/docs/ops/deployment/performance-and-scalability/). With a static environment, after the initial startup, there shouldn't be much going on, and shouldn't be resource intensive. An assessment should be done to determine the edge use case for this use case: ### Controller Otherwise we need a light weight Ingress controller: https://doc.traefik.io/traefik/providers/kubernetes-ingress/ Downside: * add package support in each package * different network components between cloud/enterprise and edge * need to satisfy mTLS/Controls elsewhere ## Logging The proposed solution for logging is to allow for Promtail/Loki/Grafana (via bigbang#647 ) to be deployed to aggregate logs into the same dashboard as prometheus metrics