Recommendations
AWS, GCP, and Azure advisors give everyone the same advice. Lens reads your accounts, your tags, your real bill and surfaces the dollars you can stop spending.
01 · RecommendationsRecommendations
Production has to stay on. Over-provisioned doesn’t.
AWS, GCP, and Azure advisors give everyone the same generic advice. Lens reads your accounts, your tags, and your real bill, 450+ rules across three clouds, every finding reconciled against what you actually pay before it reaches your dashboard.
- 450+ rules, eight categories, three clouds, cloud-native and Kubernetes
- Reconciled against live billing, not list price
- Production-classified by default
- Confidence-scored, complexity-labelled, evidence in every drawer
- 60 rules wired for remediation, apply in one click or follow the runbook
Cover what native advisors miss.
AWS Trusted Advisor, Azure Advisor, and GCP Recommender ship limited rule sets, and the cloud they run on profits when you spend more. Lens runs 450+ rules across all three, organised by idle, right-sizing, schedule, orphan, compliance, discount, security, and reliability. Every finding ships with the metric, the threshold, the action, and the dollar.
No agentic AI. No black-box ML.
Trust the dollar.
Every recommendation reconciles against your actual billing data, AWS Cost Explorer, Azure Cost Management, GCP BigQuery, before it surfaces. Current cost from your invoice. Optimised cost modelled from the same data. Azure costs amortised so reservation savings show per-resource, not as $0 at subscription scope. If the dollar isn’t on your bill, it isn’t on your dashboard.
Stay production-safe.
Destructive remediation requires admin approval before it runs. Spot conversion restricted to verified dev/test. Customer databases are explicitly excluded from mutation, RDS, Aurora, Cloud SQL, ElastiCache, Azure SQL, Postgres, MySQL never touched.
The recommendation drawer.
Four questions, answered before you act. Every recommendation tells you how much you save, how sure we are, how hard it is, and what the resource actually is. The Evidence box shows the cloud monitoring values that triggered the rule, so you see the math, not just the verdict.
- Monthly + annual savings, side by side
- Confidence score, with High / Medium / Low fallback
- Effort tag, Commitment, Spot, Migration, or Hybrid
- Generated / Re-evaluated footer timestamps
See what breaks before you apply.
Blast radius, on the canvas. Before you apply a recommendation, click View Blast Radius. zopnight maps the connected resources, shared security groups, routing targets, attachments, triggers, and classifies each one.
- Three states, affected, warning, or safe
- Risk score 0–100, weighing impact, environment, and dependents
- Impact summary before any delete, stop, or modify
Apply, audited.
60 rules wired end-to-end. 32 apply in one click; 28 are guided with confidence and complexity up front. A four-step wizard runs every time: precondition → approval → cloud action → validate.
- Destructive ops gate on admin approval
- Customer DBs excluded, RDS, Aurora, Cloud SQL, ElastiCache, Azure SQL, Postgres, MySQL
- Numbered runbook in the drawer when a rule isn’t wired for one-click
“Native advisor surfaced 8 findings. zopnight surfaced 287. Recovered $14,820/month, validated against EA-adjusted billing. Payback in under 5 months.”Internal benchmark, anonymised at customer’s request Read the methodology
