CODEX ENTRY 01INTEGRATION DECISION PATTERN · WORKED WITH / ANALYSED

Webhook vs Polling

Two ways for one system to learn that something changed — and how to decide which pattern fits the situation.

Hiring lens

Explain a technical trade-off clearly enough to decide.

Start with the problem, explain the mechanism simply, then surface the trade-offs and failure modes that change the choice.

Best fit
System Analyst · Technical BA · Integration Analyst
Verified bridge
Octopod webhook requirements · 50+ tickers · sub-second target latency
Boundary
The explainer is not a claim that every example below was production experience.
01 Mechanism02 Trade-offs03 Failure modes04 Choice
Open verified webhook evidence →
The problem first

How does one system learn that something changed in another?

There are two common patterns: either the first system keeps checking, or the second system sends a message when the change happens.

POLLING

“I’ll keep asking.”

System A checks System B every so often: “Anything new?”

Everyday example: refreshing a delivery-tracking page yourself.
WEBHOOK

“Tell me when it happens.”

System B sends a message to System A when the event happens.

Everyday example: receiving a notification when the courier arrives.
Short version: Polling = ask repeatedly. Webhook = be notified.
See the difference

One change, two ways to learn about it

System AWaits until the next check.
→
Check“Anything changed?”
→
System BReturns the latest state.
Polling is simple to control, but it keeps sending checks even when nothing changed.
Practical choice

Which one would I use?

⏱

If updates do not need to be immediate

Checking every few minutes may be perfectly enough.

Polling
⚡

If another system should react quickly

A notification when the event happens is usually a better fit.

Webhook
🛡

If missing an update would be costly

Sometimes webhooks handle fast updates while periodic checks are kept as a safety net.

Both
One level deeper: what can go wrong with webhooks?

A webhook is still just a request between systems. It can fail, arrive late, or be sent more than once.

In real integrations, the receiving side often needs to handle retries and duplicate events so that the same business action is not performed twice.

That is why “the source sent a webhook” is not the same as “the business operation definitely happened exactly once.”

See reconstructed API / idempotency evidence →

Technical wording, if you want it

Polling: the client sends scheduled requests to read the latest state.

Webhook: the source sends an HTTP request to a predefined receiver when an event occurs.

The receiver may need validation, monitoring, retry handling and duplicate protection, depending on the business operation.

Next move

Need this kind of systems thinking on your team?

If the role needs someone who can translate technical mechanisms into clear, testable decisions, let’s talk.