How exchanges handle outages and halts

An exchange outage is any period when a trading venue cannot fully process orders — because of a technical failure, an overload, or a deliberate pause. A halt is a deliberate, usually temporary stop of trading in one or more markets. Crypto exchanges run 24/7, so how they handle these events shapes how the market behaves in exactly the moments when it is most stressed.

This article explains the mechanics of outages and halts and the risks they create. It is educational only and is not trading advice.

Why outages happen at the worst moments

Exchange systems are engineered to process enormous order volumes, but demand is not constant. During sharp price moves — often driven by news or a liquidation cascade — order rates, cancellations, and data requests can spike far above normal. Common failure points include:

• Matching-engine overload. The core system that pairs buyers and sellers has finite throughput. A flood of orders can create latency or force protective throttling.
• API and website saturation. Even if the engine keeps running, the interfaces traders use to reach it can slow or fail under load.
• Infrastructure faults. Data-centre issues, network problems, or a failed software deployment can take services down independent of market conditions.

The unfortunate pattern is that outages tend to cluster around volatility, precisely when the ability to manage a position matters most.

Planned halts versus unplanned outages

Not every stop is a failure. It helps to separate deliberate halts from involuntary outages:

Traditional stock markets use formal, rule-based circuit breakers that pause trading market-wide when prices move beyond set thresholds; these are documented by regulators such as the U.S. Securities and Exchange Commission. Most crypto venues do not run a single coordinated circuit breaker, though some apply per-contract price limits, “price bands,” or their own halt rules. Because there is no market-wide standard, behaviour differs a lot from venue to venue.

What happens to your orders during a stop

The mechanics that matter most are what happens to open orders, positions, and prices while trading is interrupted:

• Open orders may not fill. If the engine is down, new market and limit orders may be rejected or queued, so you cannot rely on getting in or out.
• Existing positions still carry risk. On derivatives, a position’s mark price and margin can keep updating from an index even if you cannot trade, so liquidation risk does not simply pause because the interface is down.
• Cancellations may fail too. An outage can prevent you from cancelling or modifying resting orders, leaving them exposed when trading resumes.
• Reopening can gap. When a halted market restarts, the first available price can be far from the last one, because pent-up supply and demand clear at once.

Responsible venues publish policies for these situations — for example, how liquidations are handled during downtime, or whether certain orders are cancelled on restart. The details live in each exchange’s own terms and status documentation, and they vary.

Deposit and withdrawal pauses are their own category

Trading can keep running while the movement of funds is frozen, and vice versa, so it is worth treating deposit/withdrawal pauses separately. Exchanges pause transfers for several mechanical reasons:

• Blockchain congestion or reorganisation. If the underlying network is congested, or a chain reorganisation or fork is in progress, an exchange may pause deposits and withdrawals to avoid crediting transactions that could later be reversed.
• Wallet or infrastructure maintenance. Hot-wallet upgrades or node migrations can require a temporary transfer freeze even while trading continues normally.
• Suspected security incident. If a venue suspects an exploit, halting withdrawals is a protective measure to prevent funds leaving while the situation is assessed.

For users, the important point is that a withdrawal pause means your assets are temporarily immobile on that venue, regardless of what the market is doing. A prolonged, unexplained withdrawal freeze is also one of the more serious warning signs about a venue’s health, which is why clear communication during such pauses matters so much.

Why 24/7 markets make this harder

Traditional exchanges close overnight and at weekends, which gives operators natural windows to deploy upgrades, run maintenance, and let risk reset. Crypto venues never close. That has real operational consequences: maintenance must happen on a live market, there is no daily “reset” to clear stuck states, and a failure at 3 a.m. on a weekend still lands in the middle of active global trading. It also means staffing and monitoring have to be continuous. The absence of a close is convenient for traders but removes the built-in pauses that traditional markets rely on to stay healthy, placing more weight on the venue’s engineering and incident response.

How well-run exchanges prepare

• Capacity headroom and load testing. Building the matching engine and APIs to handle multiples of normal peak load, and testing against simulated surges.
• Graceful degradation. Shedding non-critical features (charts, secondary endpoints) to keep core order processing alive under stress.
• Status transparency. A public status page and clear incident communication so users know whether the problem is the venue or their own connection.
• Documented halt and liquidation policies. Pre-defined, published rules for how positions and orders are treated during downtime, so behaviour is predictable rather than improvised.
• Post-incident reviews. Publishing what failed and what was changed, which is a strong signal of operational maturity.

The risks these events create for users

• Loss of control. Being unable to close or adjust a position during a fast move is one of the most cited frustrations in outages, and it can convert a manageable situation into a loss.
• Liquidation while locked out. Because margin math can continue during downtime, a position can be liquidated in a window when the user could not act.
• Reopening gaps. Trading that resumes far from the last price can crystallise large moves instantly.
• Uneven information. During an incident, some users or systems may regain access before others, creating a temporary imbalance.
• Concentration risk. Relying on a single venue means a single outage can lock away both your ability to trade and your ability to move funds at the same time.

Because these risks are operational rather than about price, they are easy to overlook until an incident happens. Yet they are precisely the risks that separate a mature, well-engineered venue from a fragile one, and they become visible only under stress — which is why a venue’s track record during past incidents is worth more than its marketing during calm periods.

None of this is a reason to trade or not trade; it is a reminder that operational reliability is a real, material property of a venue, not a footnote.

What this means

Outages and halts are where a trading venue’s engineering meets the market’s worst moments. A halt can be a deliberate, protective pause; an outage is an involuntary failure — and both can leave you unable to act while your risk keeps running in the background. Because crypto has no market-wide circuit-breaker standard, how each venue prepares, communicates, and documents its halt and liquidation policies varies widely, and that operational quality is worth understanding before it is tested. For the risk mechanics that make downtime dangerous, see how liquidations cascade and how perpetual futures keep updating margin from an index price.