People ask me what tools I actually use
Not the tools I recommend in theory. The ones I actually open when I get access to a new account
AWS Cost Explorer: first stop. Not for the totals. For the service breakdown and the month-over-month deltas. I'm looking for anomalies and concentration - where does the spend cluster and what changed recently
AWS Compute Optimizer: I turn this on immediately if it isn't already on. Let it run for a couple of weeks minimum. The EC2 and RDS recommendations are genuinely useful when the utilization data is mature enough
CloudWatch Container Insights and Application Insights: for accounts running EKS or ECS, this is where I look for over-provisioned workloads and underutilized nodes. Pod resource requests versus actual usage is almost always worth examining
AWS Cost and Usage Report exported to Athena: for deep analysis. When the Cost Explorer view isn't granular enough, CUR gives you the full breakdown. I write queries against it to find specific patterns - resources with no tags, services with unexpected cost spikes, cross-region data transfer by source and destination
For GCP: billing export to BigQuery is the equivalent. It's the only way to get the full granularity. The billing console is good for overview. BigQuery queries are how you find the specific things
INFORMATION_SCHEMA.JOBS in BigQuery: for any account with significant BigQuery spend, this is mandatory. Shows you exactly which queries are processing the most data. The top 10 by bytes processed is always revealing
GCP Recommender: useful for Compute Engine rightsizing and idle resource detection. Not as mature as AWS Compute Optimizer but getting better
For both platforms: a simple spreadsheet
Every finding goes in with resource ID, current cost, projected cost after fix, effort required, risk level, and owner. This is not fancy. It's just organized. And organized information is what turns a list of findings into an executable remediation plan
The tools find the opportunities. The spreadsheet determines the order
