Zaysa

FINOPS · AUTOSTOPPING

Stop paying for servers nobody is using.

AutoStopping watches the compute you choose and stops it when CPU and network have both been quiet for your idle window. Start it again with one action, and see the savings it made.

AutoStopping on Pro
Start freeBook a demo

The usual way

Staging runs all night, and all weekend.

Development and test servers run around the clock.
A stop script on a cron job breaks and nobody notices.
A server is stopped while someone is still using it.
Nobody knows how much stopping actually saved.
What you are seeing
  1. 01Choose what it may stop
  2. 02Both signals must be quiet
  3. 03The idle clock runs
  4. 04Stopped, and confirmed
  5. 05Unclear? Do nothing.
  6. 06Back with one action
  7. 07Savings you can trust

Example data.

How it works

Choose, watch, stop

01

Choose resources and an idle window

Pick a cloud account, the resources this rule covers and an idle threshold in minutes.

staging nights and weekends

AWS · shop-staging

staging-api · EC2 m6i.large

staging-workers · Auto Scaling group, 3 × t3.large

Idle threshold · 30 min

Example data.

02

Watch CPU and network

A resource is idle only while both signals stay below their thresholds. Rules are checked every 5 minutes.

staging-api · checked every 5 minutes

✓ CPU 1.2%
below 2%

✓ Network 18 KB/s
below 50 KB/s

Example data.

03

Stop, then start with one action

The rule stops chosen resources after the idle window. Use Start on the rule when you need them again.

staging-api

active → cooling → stopped

Power state: stopped · read live from AWS

Start with one action on the rule.

Example data.

Idle detection

Idle means really idle

Move the CPU and network readings. Both must stay below their thresholds for the idle window before a stop. When network cannot be read, the CPU threshold is stricter.

staging-api · live readingIdle

When network cannot be read, CPU must be below 1%.

Example data.

Supported compute

Only the compute you choose

A rule names one cloud account, the resources it covers and an idle threshold in minutes. Only resources you put in a rule are ever stopped.

AWSEC2 instances · Auto Scaling groups
AzureVirtual machines · deallocated when stopped
DigitalOceanDroplets · powered off when stopped

Example data.

Rule detail

Every stop on record

Each rule shows Idle time, Compute, Savings and Last activity. Its detail drawer shows cumulative savings, spend versus savings for the last seven days and state changes. Savings use the latest cost report; when cost is unknown, they stay at 0.

Rule detail

staging nights and weekends
stopped

Cumulative savings

$1,284.60

Last activity

Stopped · 30 min idle
Spend vs savings (last 7 days)

■ Spend   ■ Savings

State changes

active → cooling · quiet for 30 min

cooling → stopped · read live from AWS

stopped → warming → active · Start

Idle time · 30 minCompute · staging-apiSavings · from cost reportLast activity · Start

Example data.

On the Pro plan

Stop the waste you can see coming

Put a quiet idle window around the compute you choose, then see when it stopped and what it saved.

AutoStopping

Stop chosen compute after a quiet idle window.

Waste Cleanup

Review priced waste findings and record cleanup.

Budgets and alerts

Watch account budgets and cost alerts.

Commitments

Review coverage and utilisation.

$20 per month

See the Pro planBook a demo

Questions

Frequently asked questions

AWS EC2 instances and Auto Scaling groups, Azure virtual machines, and DigitalOcean Droplets. Azure virtual machines are deallocated when stopped; DigitalOcean Droplets are powered off. Only resources you put in a rule are ever stopped.

More in FinOps

Waste CleanupCloud HistoryAll features

See what you run, and what it costs.

Free plan, no card needed.

Start freeBook a demo