Overview

Live fleet & throughput telemetry.

Workers alive

Queue depth

jobs waiting

Success rate

Burn rate

Active workers

Open fleet →

Fleet

Cloud workers — each ~one server solving captchas.

Scale to

Create workers — pick a plan to A/B cost vs. speed (autoscaler off while mixing plans)

Plan

Count

Threads / worker

Est:

Fleet config

Snapshot

Server catalog

Queue

Test solve

Waiting jobs

Task	URL	Sitekey	Age

Queue empty.

In-progress by worker

Autoscaler

Scaling cockpit & policy.

Current → Desired

Effective max

Next tick

Session left

Last action:

Scaling policy

Autoscale enabled

Pause scale-up

Pause scale-down

Min workers

Max workers

Queue / worker

Idle scale-down (s)

Worker cost €/hr

Cost cap €/hr

Session max (min)

Session

Scaling timeline

Solver config

Schema-driven solver parameters.

Parameters

Presets

Config snapshots

Worker provisioning

Threads / worker · concurrent solves — keep LOW (1 proven; higher starves CPU & tanks solve rate)

Baked into each NEWLY-booted worker (running workers keep their boot value). For throughput, scale OUT — more workers, not more threads.

Cost & budget

Spend tracking and limits.

This run

Today

This month

Projected (session)

Daily spend

Budget limits

Daily €

Monthly €

Hard stop €

Soft warn %

Alerts

Threshold rules & firing history.

Rules

History

Audit log

Administrative actions trail.

Access & keys

API key inventory & posture.

Prefix	Label	Role	Scopes	Last used

No keys.

Health

Infrastructure & self-test.

System

Fleet snapshot

Ops

Deploy actions (admin). 501 = ops disabled.

Self-test

Live logs (SSE)

Solves

Solve history, traces & latency.

Latency (s)

By challenge type

Time	Task	Result	Type	Worker	Exit IP	Elapsed	Model	Reason

No solves.

Solver Control Center