> ## Documentation Index > Fetch the complete documentation index at: https://docs.nofire.ai/llms.txt > Use this file to discover all available pages before exploring further. # Cost Optimization > Find where you're overspending on infrastructure and observability tooling. NOFire helps you find waste in two areas: your **infrastructure spend** and your **observability tooling costs**. **Typical impact:** 20-35% savings identified on observability tooling. 15-25% AWS cost reduction on first pass. *** ## Infrastructure Costs See costs by service, team, environment, or region. No BI tools needed. When costs jump unexpectedly, NOFire identifies what changed and which resources are responsible. Find EC2, RDS, and other resources that cost more than their actual usage justifies. Rightsizing recommendations included. Unattached EBS volumes, idle load balancers, orphaned resources that are costing money doing nothing. Instance utilization analysis showing daily cost and efficiency scores by resource category ### How to use it ``` @NOFire AI what was our AWS spend last week, broken down by service? ``` ``` @NOFire AI our costs spiked $12,000 last Tuesday — what changed? ``` ``` @NOFire AI which RDS instances have low utilization but high cost? ``` ``` @NOFire AI find unused resources — unattached EBS volumes, idle load balancers ``` ``` @NOFire AI show me our top 10 cost drivers this month vs last month ``` *** ## Observability Costs Prometheus, Grafana Cloud, Elastic, and similar tools charge by volume. NOFire analyzes your actual telemetry usage and finds what you're paying for but not using. **Typical impact:** 20-30% of observability spend is optimizable overhead. Metrics that are collected and stored but never queried in any dashboard or alert rule. Label combinations that explode your time series count and drive up costs. Services producing excessive log volume: debug logs in production, verbose libraries, redundant entries. Where you're over-sampling (paying for traces nobody looks at) or under-sampling (missing issues). ### How to use it ``` @NOFire AI which metrics are we collecting but never using in any dashboard or alert? ``` ``` @NOFire AI which labels are causing high cardinality in our Prometheus? ``` ``` @NOFire AI which services are generating the most log volume and is it necessary? ``` ``` @NOFire AI how much could we save by removing unused metrics from Grafana Cloud? ``` *** ## Common Scenarios ``` @NOFire AI we're paying too much for our observability stack — where's the volume coming from? ``` NOFire checks which metrics, log sources, and trace volumes are driving the bill and which you could cut without losing visibility. ``` @NOFire AI our Prometheus is running out of memory — which metrics have the highest cardinality? ``` ``` @NOFire AI give me a summary of our observability spend — what are we paying for vs what we use? ``` ``` @NOFire AI our AWS bill jumped 30% this month — break it down ``` ## Available in * **Slack**: `@NOFire AI` in any channel * **Chat dashboard**: [my.nofire.ai](https://my.nofire.ai) * **IDE (MCP)**: quick cost checks while coding ## Getting Started For infrastructure costs. Requires read-only IAM access and Cost Explorer enabled. [Connect AWS →](/providers/aws) For observability cost analysis. [Connect Grafana →](/monitoring/grafana) Start broad: `@NOFire AI where are we overspending?`