Failover for a trading bot means that when one link in the chain fails, a second path takes over or you recover quickly without doubling or losing a trade. For a TradingView trader the chain has three links: the alert, the webhook receiver and the exchange API. The simplest way to harden it is to run the middle link on a hosted platform instead of a machine you keep alive yourself. Tickerly turns TradingView strategy alerts into live orders with no VPS and no coding, so the failure point you would otherwise own is not yours to maintain.
Key Takeaways
- A TradingView bot fails in three places: the alert, the receiver and the exchange connection. Plan failover for each one separately.
- Failover is only safe if recovery never blindly repeats an order. After any outage, check the exchange’s real positions first.
- TradingView can send the same webhook up to four times after certain server errors, so duplicate protection matters as much as uptime.
- Test recovery on an exchange demo account, never first with live funds.
- A hosted, no-code receiver removes the server you would otherwise have to keep running and replace.
Table of Contents
- What failover means for a TradingView-connected bot
- The three links that can break
- Failure map: what breaks and what you do
- The golden rule of recovery: check before you resend
- How to set up a failover-minded TradingView automation
- How to test recovery safely
- Failover checklist
- Our perspective
- Put your automation on a hosted receiver
- FAQ
- Sources
What failover means for a TradingView-connected bot
In institutional trading, failover is the automatic switch from a primary system to a standby. The FIA’s guide to automated trading systems recommends standby systems for production infrastructure and key trading services, plus business-continuity plans that are actually tested. It also stresses that recovery often depends on critical service providers outside your own application.
That last point is the useful one for you. A retail TradingView bot does not run its own matching engine or data center. It depends on three outside services: TradingView to fire the alert, a receiver to translate it, and an exchange or broker to fill it. Failover for you is therefore not about building redundant servers. It is about knowing which link broke, how to tell, and how to get back to a correct position.

A note on regulation: rules such as the FCA’s MAR 5A.5 controls for certain UK trading venues, or the MiFID II Article 17 requirements for EU investment firms engaged in algorithmic trading, apply to those venues and firms. They are not universal legal duties for every individual running a bot. They are still a good template for what a continuity plan covers: resilience, tested systems and recovery from system failures.
The three links that can break
1. The alert
The chain starts when your Pine Script strategy or indicator condition triggers an alert in TradingView. If the alert is paused, expired, or was created on the wrong alert type, nothing downstream happens. Alerts that never fire are the quietest failure because nothing errors. Our guide to missing TradingView alerts covers why they vanish, and the difference between strategy alerts and study alerts matters when you decide what to automate.
2. The receiver
The webhook receiver is the service at the URL your alert calls. If it is down or returns a server error, the order is not placed. This is where a self-run setup is most fragile: a single server, a single process, a certificate or a disk that fills up. TradingView’s documentation says that when a receiver answers with an HTTP 500 to 599 error other than 504, it resends after 5 seconds, up to 3 more times, for 4 sends in total.
That behavior cuts both ways. It gives a short outage some room to recover. It also means the same alert can arrive more than once, so any recovery path has to handle repeats. We go deeper in webhook retries for trading bots and how to prevent duplicate orders.
3. The exchange connection
The last link is the exchange API. Exchanges rotate and limit connections, and they have maintenance windows. For example, Binance’s Spot WebSocket API documentation states that connections are limited to 24 hours and clients should handle disconnection and reconnection. It also describes a server ping every 20 seconds, with the connection dropped if no pong arrives within a minute. Those details concern direct API clients, but they show a general truth: connections to exchanges end, and a design that assumes they never do will eventually miss a trade.

API keys are part of this link too. A key that was revoked, expired or lost its trading permission looks, from the outside, exactly like an outage. Our exchange setup guide shows how the connection is made, and the list of supported exchanges tells you where Tickerly connects.
Failure map: what breaks and what you do
| Failure | What you see | Recovery action |
|---|---|---|
| Alert paused or expired in TradingView | No signal, no order | Re-enable the alert, check alert limits |
| Receiver unavailable | Alert fired, no order | Wait for TradingView’s resends, then verify positions |
| Same alert delivered several times | Possible repeat orders | Compare exchange positions to what the strategy expects |
| Exchange API key invalid | Orders rejected | Recreate the key with trading permission |
| Exchange maintenance or disconnect | Orders delayed or rejected | Confirm state on the exchange, then resume |
| MetaTrader terminal closed | MT4/MT5 orders do not execute | Restart the terminal and verify open positions |
The golden rule of recovery: check before you resend
The most expensive failover mistake is resending an order whose first attempt actually succeeded. A timeout does not tell you the order failed. It tells you that you did not receive an answer. The order may have filled while the response was lost.
A fail-safe pattern documented in an open-source reconciliation module for trading agents shows the logic well: after a crash or timeout, each unresolved order is matched against the venue’s live orders. An order that is not found stays blocked, because it may have filled and closed, and sending it blind would double the trade. Only an explicit human acknowledgement clears it.
You can apply the same principle by hand as a TradingView trader:
- After any outage, open the exchange and look at actual open positions and recent fills.
- Compare them to what your strategy believes it holds in TradingView.
- Fix any mismatch manually on the exchange before the next alert fires.
- Only then let automation continue.
If you want to understand why the order’s status on the book matters more than your last message, this primer on order book basics explains how resting orders and fills appear on a venue.
In Tickerly, position sizing, stops and risk rules live in your TradingView strategy, not in the platform. That is why the reconciliation step compares the exchange with your strategy’s logic: the strategy defines what the position should be, and the exchange tells you what it is.
How to set up a failover-minded TradingView automation
Step 1: Use a receiver you do not have to babysit
A receiver on a single home computer or a rented VPS is a single point of failure that you must patch, monitor and restart. Tickerly is hosted and needs no VPS or coding, which takes that whole layer off your list. You connect your exchange, point your TradingView alerts at it, and the setup is the same whether you run one strategy or many. Our walkthrough of TradingView strategy automation shows the path from alert to live order, and TradingView alerts to exchange covers routing.
Execution within a fraction of a second after an alert matters here too. A short delay on a fast-moving market is a failure in its own way, and we explain how to think about it in real-time trade execution.
Step 2: Spread risk across exchanges and strategies
A second kind of failover is operational: do not let one account carry everything. Tickerly supports multiple strategies, assets and platforms at once, so you can split strategies across exchange accounts. If one exchange has maintenance, the others keep working. This is diversification of infrastructure, not a promise about trading outcomes.
Step 3: Keep your alert capacity in order
A failover plan that needs one more alert and cannot create it is useless. Check your TradingView plan’s alert limits ahead of time. Our guides on TradingView alert limits and choosing a TradingView plan help you size it.
Step 4: If you trade MetaTrader, plan for the terminal
When you route TradingView alerts to MT4 or MT5, your MetaTrader terminal must be running for orders to execute. Treat that terminal as part of your failover plan: know where it runs, how you restart it and how you confirm positions after a restart. The TradingView to MT5 guide walks through the connection.
Step 5: Secure the path
A hijacked alert URL is also an outage, because it can cause unwanted orders and force you to shut down. Read TradingView webhook security and keep your exchange API keys limited to the permissions the bot needs.
How to test recovery safely
Institutions rehearse failure. The FIA’s guide recommends validating contingency plans with exercises and tabletop reviews and reviewing them regularly. The ASX describes test environments that simulate order-entry failover and failback between primary, secondary and tertiary sites. You do not need that scale, but you can copy the habit.
Testing uses the exchange’s demo account, since Tickerly has no paper-trading mode of its own. A practical drill:
- Connect a demo account at an exchange that offers one and run your alert through the full chain.
- Pause the alert in TradingView, then fire a test condition and confirm that nothing reaches the exchange. This proves your alert-level monitoring works.
- Revoke the demo API key, trigger an alert and see how a rejected order looks to you. Then recreate the key and confirm trading resumes.
- Reconcile. After each step, compare the demo positions to what your strategy expects. Write down any mismatch and how you fixed it.
- Repeat quarterly, and after any change to your strategy, exchange or API keys.
For the monitoring side of a drill, a low-risk migration plan for observability is a good model of how to stage a controlled switch, validate recovery and keep a rollback path.
Failover checklist
- Alerts are active and not near expiry in TradingView.
- Alert capacity covers your strategies plus spares.
- The exchange API key has trading permission and nothing more than needed.
- You know how to find your open positions on the exchange in under a minute.
- Your strategy’s expected position is written down, so mismatches are obvious.
- You have rehearsed a recovery on a demo account.
- For MT4/MT5, you know how to restart your terminal.
Our perspective
Most traders think about failover as a server problem. We think of it as a state problem. Servers come back. What hurts is not knowing whether the last order went through.
That is why we built Tickerly as a hosted, no-code bridge between TradingView and your exchange. You do not run a VPS, you do not maintain a webhook server, and you can run unlimited strategies and alerts across several exchanges from one place. The fewer moving parts you own, the fewer places a failure can hide.
Our advice is simple: keep your risk rules in your TradingView strategy, keep your exchange as the source of truth for positions, and rehearse recovery on a demo account before it matters. Automation should make your trading calmer, not leave you guessing.
Jay
Put your automation on a hosted receiver
If you want fewer links to maintain, Tickerly connects your TradingView alerts to exchanges for crypto, forex, stocks, futures and prop firm markets. Plans start at $19 per month with Level 2, with Level 3 at $29 and Level 4 at $39, and every plan starts with a 30-day free trial. Compare them on the Tickerly plans page.
The alert log is a good place to start when you want to see what reached the platform after an outage.
FAQ
Do I need a VPS for trading bot failover?
Not with Tickerly. It is hosted and requires no VPS or coding, so there is no server of yours to fail over. If you trade through MetaTrader, your own MetaTrader terminal must be running, so plan for that machine separately.
Can a TradingView alert create duplicate orders?
It can be delivered more than once. TradingView resends a webhook up to three more times, four sends in total, when the receiver returns certain server errors. After any outage, compare your exchange positions with your strategy before you continue. See our guide on preventing duplicate orders.
What should I do first after an exchange outage?
Open the exchange and check open positions and recent fills before anything else. A timed-out order may still have filled. Fix any difference manually, then let alerts resume.
How do I test failover without risking money?
Use the exchange’s demo account. Run an alert through the full chain, pause alerts, revoke the demo key and recover, and reconcile positions after each step.
Sources
- FIA, Guide to the Development and Operation of Automated Trading Systems (March 2015)
- FCA Handbook, MAR 5A.5 systems and controls
- ESMA, MiFID II Article 17 Algorithmic trading
- TradingView, Webhook resubmission
- Binance Spot WebSocket API documentation
- agentic-trading-mcp reconcile module
- ASX, Guide to ASX 24 Testing Services
- Order book basics
- Low-risk observability migration plan
Recommended
- TradingView Strategy Automation: How to Go From Alert to Live Order
- Webhook Retries for Trading Bots: Avoid Duplicate Orders
- Secure Crypto TradingView Bot Setup: Avoid Duplicate Trades
- How to Automate Trading on TradingView Using Tickerly
This article was produced with Al assistance and reviewed for accuracy. It is provided for general information only and is not professional or financial advice.

