Token Usage & LLM Cost
Terraform Generation & Deployment Status
Resources & Cost
Ansible Generation & Execution
Revenue & Margin (modeled — LLM unit economics)
Modeled, not billed. Revenue = resources added × $1/resource (a pricing assumption, not invoiced revenue). Cost = LLM token cost only, all sources (no cloud/infra COGS). “Margin” here is a token-efficiency ratio (revenue − LLM cost), not an accounting profit margin — don’t use it for profitability decisions.
Over a bounded date range these figures are approximate: revenue is anchored on when resources were added, while cost is anchored on when tokens were spent, and the LLM cost to generate a stack often lands in a different window than the apply. They converge only over all-time.
Per-User Token & Cost
| User | Cost | Tokens | Calls |
|---|
Per-Team Token & Cost
| Team | Cost | Tokens | Calls | Users |
|---|
Per-User Terraform & Deployment Detail
Generated = all Terraform code produced. Apply outcome is event-based: a plan that was applied and later destroyed still counts as Fully Applied (a destroy implies a prior apply), so applies are never hidden by a later teardown. Deployments = plans that actually ran an apply (Fully + Partially + Failed) = Generated minus Not Applied. Destroyed is an overlapping lifecycle flag (mostly a subset of Fully Applied), not part of that sum. New Stacks = distinct infrastructure stacks first created in the period, so a high Deployments / low New Stacks (or low Added) row just means heavy re-apply/iteration on pre-existing stacks, not missing resources.
Per-user money is approximate. Token Cost is attributed to the user who ran the LLM calls (user_id), while resources are attributed to the stack’s creator (created_by) — upstream these can be different people (plan owner vs. first applier). Aggregate totals are exact; per-user Token Cost / Avg Cost/Resource can misattribute between users. Deleted is also a lower bound (some modify-apply deletes aren’t state-diffed upstream).
| User | Generated | Deployments | New Stacks | Applied | Partially Applied | Failed | Destroyed | Not Applied | Apply Success Rate | Resources | Added | Deleted | Token Cost | Avg Cost/Resource |
|---|
Per-User Ansible Playbook & Execution Detail
| User | Playbooks Generated | Not Executed | Executions | Succeeded | Failed | Success Rate | Avg Execution Time |
|---|
Cost Metrics (Token Cost per Resource Deployed)
Average = total token cost ÷ total resources deployed for the period. Min / Max / Median / P25 / P90 / P99 are computed across each day's own cost-per-resource ratio, to surface outlier days the blended average can hide. Note: each day is weighted equally regardless of volume, so a day with only 1–2 resources can dominate Min / Max / P99. The Average (volume-weighted) is the robust figure; read the percentiles as day-shape, not resource-level, statistics.
Platform Activity Overview
Messages, Terraform Generated, Token Cost & Resources Deployed (Indexed Trend)
Each series is scaled to % of its own period maximum so trends of very different magnitudes (message counts vs. dollars vs. resource counts) can be compared on one axis.
Terraform Generated vs. Token Cost vs. Resources Deployed
Daily Relationship (bubble size = resources deployed that day)
Terraform Lifecycle vs. Token Cost
Deployed / Re-applied / Deleted Events vs. Token Cost Over Time
Token Usage & Cost Detail
Tokens & Cost Over Time
Cost by Model
Cost by Agent / Source
Stacks by Status Over Time
Stacks by Status
Terraform Apply Outcome
Resources & Cost by Cloud Provider
| Provider | Resources | Token Cost | Avg Cost/Resource |
|---|