Recherche vérifiée par sources: Binance / Kraken

Verified source facts

1. A platform upgrade can affect authentication and transfers while core trading continues. 2. Binance identifies an at-least-three-hour upgrade window and lists affected service groups. 3. Kraken publishes component-specific status notices for API and funding functions. 4. A general HTTP 200 or usable homepage does not prove that every service plane works.

Comparison at a glance

1. Control test: Login

A trading-control disclosure is meaningful only when converted into a testable rule. For login, inspect session creation. Use a sandbox, non-production context, or the smallest reversible test available, then test separately. Capture configuration and event evidence without recording credentials, and verify the documented cancel, reject, decrement, or history behavior.

Model the failure path, not just the setup path. Ask which strategy owns the maker and taker orders, which account boundary the venue evaluates, which message reveals intervention, and how reconciliation behaves after a partial fill or disconnect. Platform documentation describes supported settings; it cannot prove that an operator configured them correctly. In this dimension, access failure is therefore an explicit audit gap.

The login row should end with a concrete monitoring rule tied to session creation. Make the operator test separately, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; access failure remains a separate production risk until observed.

2. Control test: Registration

For registration, inspect new account flow. Use a sandbox, non-production context, or the smallest reversible test available, then defer if unstable.

In this dimension, duplicate submission is therefore an explicit audit gap.

The registration row should end with a concrete monitoring rule tied to new account flow. Make the operator defer if unstable, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; duplicate submission remains a separate production risk until observed.

3. Control test: Authentication

For authentication, inspect 2FA and MFA. Use a sandbox, non-production context, or the smallest reversible test available, then keep backup.

In this dimension, lockout is therefore an explicit audit gap.

The authentication row should end with a concrete monitoring rule tied to 2FA and MFA. Make the operator keep backup, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; lockout remains a separate production risk until observed.

4. Control test: Trading UI

For trading ui, inspect manual order surface. Use a sandbox, non-production context, or the smallest reversible test available, then verify live.

In this dimension, false availability is therefore an explicit audit gap.

The trading ui row should end with a concrete monitoring rule tied to manual order surface. Make the operator verify live, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; false availability remains a separate production risk until observed.

5. Control test: Order API

For order api, inspect REST, WebSocket or FIX. Use a sandbox, non-production context, or the smallest reversible test available, then monitor component.

In this dimension, automation failure is therefore an explicit audit gap.

The order api row should end with a concrete monitoring rule tied to REST, WebSocket or FIX. Make the operator monitor component, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; automation failure remains a separate production risk until observed.

6. Control test: Deposits

For deposits, inspect credit path. Use a sandbox, non-production context, or the smallest reversible test available, then retain transaction ID.

In this dimension, balance delay is therefore an explicit audit gap.

The deposits row should end with a concrete monitoring rule tied to credit path. Make the operator retain transaction ID, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; balance delay remains a separate production risk until observed.

7. Control test: Withdrawals

For withdrawals, inspect approval and broadcast. Use a sandbox, non-production context, or the smallest reversible test available, then avoid urgency.

In this dimension, stuck request is therefore an explicit audit gap.

The withdrawals row should end with a concrete monitoring rule tied to approval and broadcast. Make the operator avoid urgency, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; stuck request remains a separate production risk until observed.

8. Control test: Transfers

For transfers, inspect internal ledger movement. Use a sandbox, non-production context, or the smallest reversible test available, then pre-position funds.

In this dimension, margin shortage is therefore an explicit audit gap.

The transfers row should end with a concrete monitoring rule tied to internal ledger movement. Make the operator pre-position funds, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; margin shortage remains a separate production risk until observed.

9. Control test: Payments

For payments, inspect merchant or pay rail. Use a sandbox, non-production context, or the smallest reversible test available, then check status.

In this dimension, duplicate payment is therefore an explicit audit gap.

The payments row should end with a concrete monitoring rule tied to merchant or pay rail. Make the operator check status, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; duplicate payment remains a separate production risk until observed.

10. Control test: Recovery

For recovery, inspect official resolved state. Use a sandbox, non-production context, or the smallest reversible test available, then run small test.

In this dimension, premature resumption is therefore an explicit audit gap.

The recovery row should end with a concrete monitoring rule tied to official resolved state. Make the operator run small test, retain the venue response, and assign an owner for exceptions. Documentation quality is a useful signal, but it does not certify the exchange or the trading system; premature resumption remains a separate production risk until observed.

Decision scenarios

Scenario 1: when login becomes decisive

Consider a reader whose main constraint is session creation. The correct response is not to infer a universal platform or product winner. It is to test separately, retain the evidence, and define a stop condition before funds or permissions are committed. If the evidence changes, the decision must be reopened. This produces a repeatable operating rule rather than a one-time opinion.

Scenario 2: when registration becomes decisive

Consider a reader whose main constraint is new account flow. It is to defer if unstable, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 3: when authentication becomes decisive

Consider a reader whose main constraint is 2FA and MFA. It is to keep backup, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 4: when trading ui becomes decisive

Consider a reader whose main constraint is manual order surface. It is to verify live, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 5: when order api becomes decisive

Consider a reader whose main constraint is REST, WebSocket or FIX. It is to monitor component, retain the evidence, and define a stop condition before funds or permissions are committed.

Scenario 6: when deposits becomes decisive

Consider a reader whose main constraint is credit path. It is to retain transaction ID, retain the evidence, and define a stop condition before funds or permissions are committed.

Sources and editorial method

• System upgrade notice — Binance Support
• Component status and maintenance — Kraken Status
• Announcement center — Binance Support