Questions platform teams actually ask
Does Titansama's collector have write access to my cluster?
+No. By default the collector ships with a read-only ClusterRole. Any automated action — scaling, rightsizing, remediation — runs through a separate, explicitly-granted, approval-gated path you control.
What data actually leaves my cluster?
+Metrics, events, and log metadata used for correlation — batched and encrypted in transit. Full log payloads stay opt-in per namespace, and you can restrict export entirely for regulated workloads.
How is this different from Prometheus + Grafana?
+Titansama doesn't replace your metrics store — it sits alongside it and adds correlation, triage agents, and cost attribution on top. Most teams keep Prometheus for raw retention and use Titansama for the "what does this mean" layer.
Can CostCurtail work without the rest of the platform?
+Yes. The FinOps module can run standalone against your billing export and cluster metrics if you're not ready to adopt fleet monitoring or the agent layer yet.
Does InfraAIOps require a single cloud provider?
+No. InfraAIOps is built for mixed estates — connect AWS, Azure, and GCP independently, and policies, identity checks, and compliance mappings apply consistently across all three.
What's the actual resource footprint?
+One Deployment per cluster, ~96Mi memory request, ~100m CPU. No per-pod sidecars and no DaemonSet required for the core monitoring path.
Stop reading dashboards.
Start reading conclusions.
Titansama is in guided rollout with a limited set of platform teams.