SIMULATIONYou are using a browser simulation, not the real lab. Every reading is generated.Install the real lab on GitHub ↗How it works ↗
SIMULATOR

Browser-only simulation. Services, failures and measurements are generated in this tab; no containers run and nothing here is real. The full lab runs real Docker services on your machine.

How it works ↗⬇ Install the full lab
SIMULATIONBefore you start

This is a simulator, not the real lab.

AegisRack is a hands-on lab that stops real Docker services so you can practise finding and fixing the failure. This page runs the same interface against a JavaScript simulation in your browser:

To use the real thing, install the full lab from GitHub. It takes about 10 minutes on Linux or Windows (WSL2) with Docker.

⬇ Install the full lab on GitHubHow it works ↗

Simulator

Your fleet. Your investigation.

Follow the evidence, test a hypothesis, and bring your services back.

ObserveInvestigateVerify

Starting the simulated collector…
Scenario 01 · Service dependencies

When a dependency misbehaves

Two simulated services. One simulated fault: stopped, starved of CPU, or out of disk. Find the boundary of the failure.

BaselineWaiting for measurements
No active timer
  1. 01Baseline
  2. 02Start
  3. 03Investigate
  4. 04Recover
  5. 05Verify
  6. 06Reset
Suggested next · optional

Connecting…

Waiting for the first measurements.

01 / Live workspace

Fleet topology 02

Simulated in your browser
LIVE MAP Drag nodes · Shift + arrows
ControllerUI · checks · evidence
BrokerDocker · inspect · start · stop
Waiting for observed requests
HealthyDegradedUnavailableUnknown / stale
Recovery is measured, never assumed.

A successful control action is followed by three healthy verification cycles.

02 / Selected node

gateway-a

INSPECT

Choose a check. Explore in any order.

Operator actions · selected node

Safe to try. An action that finds nothing to change says so; the evidence decides whether it helped.

03 / SIMULATED OBSERVATIONS

Evidence notebook 0

Open a record to inspect its result. Select evidence to support your diagnosis. Export within 24 hours to preserve detailed observations.

Following new observations
No observations yet.
Simulated checks will appear here when the collector runs.
04 / Your pace. Your path.

A little direction?

Explore unaided, or choose just the support you need. Help is recorded, never forced.

✧ AI investigatorUPCOMING

The read-only AI assistant is not implemented yet. All current guidance is deterministic. No model requests are sent.

05 / Optional · not a recovery gate

Your working diagnosis

0 evidence selected
06 / Incident replay

Incident report

No investigation yet

After an investigation starts, its latency, failures and your actions appear here. Press play to replay it.

07 / Attempt ≠ outcome

Action timeline

Recovery attempts and verification will appear here.

08 / THIS TAB ONLY

Investigation history

Your investigations will be kept here.

Local evidence storage
Session timer

Keep investigating?

Your 15 minutes are up. Continue exploring, or restore the lab to baseline.

Investigation archive

Investigation

↗ Export this investigation