What-If Simulator

An interactive scenario engine that re-runs your last 90 days of cloud usage against any hypothesis: a new Savings Plan, a 30% traffic spike, a region migration, or a switch from on-demand to Spot — all on real data.

  • 90 days of real usage
  • Stackable scenarios
  • Confidence bands
lumicost.com — what-if-simulator
SCENARIO · MIGRATE EC2 → SAVINGS PLANS · 12 MO PROJECTION30% commitmentDELTA · 12 MO−$1.86M

By the numbers

What-If Simulatorin 4 numbers

DOWNSIZE · RESERVED_INSTANCE · SCHEDULE_STOP · DELETE · STORAGE_LIFECYCLE

0

scenario types

simulation is fully read-only

0

writes to your accounts

totalEstimatedAnnualSavings

0

multi-scenario aggregation

DRAFT · RUNNING · COMPLETED · FAILED

0

simulation states

How it works

Three steps. No agents. No surprises.

1

Define scenarios

Each scenario picks a resource and one of: DOWNSIZE (newInstanceType), RESERVED_INSTANCE (1y/3y term), SCHEDULE_STOP (start/end window), DELETE, or STORAGE_LIFECYCLE (transition + delete after N days).

5 primitives · per-resource

2

Run

POST /simulations/{id}/run executes a strategy per type: DownsizeStrategy compares current vs new instance price; ReservedInstanceStrategy applies the term discount; ScheduleStopStrategy applies the off-hours fraction; DeleteStrategy zeroes the cost; StorageLifecycleStrategy combines transition + retention math.

deterministic · per-resource

3

Compare & decide

Each run returns scenarioBreakdowns[] (with insights + warnings) plus a totalEstimatedAnnualSavings — share the URL with the team, decide, then act through the recommendation flow.

share · decide · act

What's inside

An interactive scenario engine that re-runs your last 90 days of cloud usage against any hypothesis: a new Savings Plan, a 30% traffic spike, a region migration, or a switch from on-demand to Spot — all on real data.

01

Hypothesis-driven, not vibes-driven

Drag a slider for commitment coverage, pick a target instance family, dial workload elasticity. The simulator replays 90 days of usage and produces a band, not a point estimate, so you sign with eyes open.

02

Stack scenarios

Combine 'move 40% of stateless to Spot' + 'sign 60% commit on m6i' + 'consolidate 3 regions to 2' and see the joint effect — including the second-order networking cost.

03

Sharable as a snapshot

Save a simulation and send the URL to finance, security or your cloud rep — they see the same numbers without poking around your account.

Capabilities

Everything you need on day one.

  • DOWNSIZE scenario
  • RESERVED_INSTANCE 1y / 3y
  • SCHEDULE_STOP off-hours window
  • DELETE (full cost reduction)
  • STORAGE_LIFECYCLE (transition + delete)
  • Multi-scenario aggregation
  • Per-scenario insights & warnings
  • 100% read-only on your accounts

Before vs after Lumicost

Night and day.

Without Lumicost

  • Spreadsheet that conflates 5 levers into one
  • 'What if we RI 30%?' → no answer for days
  • Buying the wrong RI eats the savings

With Lumicost

  • Five explicit primitives, composable
  • Scenario in seconds with annual savings
  • Compare 1y vs 3y before you commit

Where teams use it

Three ways teams put it to work.

Pre-renewal commitment sizing

Stop signing the renewal at last year's coverage — re-baseline against the last 90 days, not the last quarter's spike.

Black Friday capacity

Model a 4× burst with realistic auto-scaling latency and find the cheapest instance mix that doesn't blow your SLO.

Region consolidation

See egress, latency and cost effects of collapsing two regions into one — before kicking off the migration sprint.

From the field

Five primitives composed cleanly will save more money than ten that nobody understands. We resisted the urge to ship a Monte Carlo for a reason.
The Lumicost teamdesign principle

Plays nice with your stack

AWSGCPAzureOracle CloudKubernetesTerraformPulumiSlackJiraLinearGitHubGitLabDatadogSplunkOpenTelemetryPagerDutyNetSuiteSAPAWSGCPAzureOracle CloudKubernetesTerraformPulumiSlackJiraLinearGitHubGitLabDatadogSplunkOpenTelemetryPagerDutyNetSuiteSAP

FAQ

Frequently asked questions

How accurate are the simulations?

Within ±5% on customers with stable workloads, ±10% during seasonal events. Confidence bands are always shown.

Can I import an external scenario?

Yes — JSON import. Useful for replaying a vendor proposal against your real footprint.

Try it free.

Connect read-only credentials. First insights in 24 hours. No credit card.

90 seconds · read-only credentials · no credit card