Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsAmazon EventBridge cost is driven by more than the number of events you publish. Event category, payload size, delivery path, account and Region boundaries, archives, replay, schema discovery, API destinations, and adjacent services can each change the bill. AWS currently lists management-event ingestion as free, while partner and custom events use paid ingestion meters; delivery and optional features follow their own rules. Verify the live Amazon EventBridge pricing page for your Region before committing to an estimate. The figures and rules below were checked on October 13, 2026.
The reliable way to forecast a bill is to map every event path, convert payloads into the applicable billing chunks, and then add optional feature and network meters. A logical event is not always one billable unit.
As an Amazon Associate I earn from qualifying purchases.
What an EventBridge bill measures
EventBridge pricing separates the event bus from features that sit around it. For each workload, identify four dimensions first:
Recommended Free Tools
- Origin: AWS management, partner, or custom events.
- Volume: events published, delivered, archived, replayed, or discovered during the billing period.
- Payload: the size distribution of events, because payloads are rounded into billing chunks.
- Route: the buses, targets, accounts, Regions, and network connections that an event crosses.
The current AWS pricing page illustrates separate paths for management events, partner events, and custom events. It also distinguishes delivery to a bus from delivery to a service and shows that crossing an account boundary can create a different delivery charge. Do not apply one per-event assumption to every source and destination.
#1 Best Overall
Event category and destination are the first cost split
AWS management events
AWS describes management-event ingestion as free on the current pricing page. That does not make every downstream action free: delivery to another account, delivery to a service, data transfer, archives, replay, or other enabled features can still be metered. Record the route after ingestion rather than stopping your estimate at the source.
Partner events
Partner events have an ingestion charge under the pricing categories AWS publishes. The destination still matters, especially when events are forwarded across accounts or Regions. If a partner event source does not have an event bus attached, use the current EventBridge FAQ to confirm the applicable billing scope instead of assuming that the absence of a bus makes the event free.
Custom events
Custom events use the paid custom-event ingestion category. Count where they are published, where rules send them, and whether a forwarding rule creates another delivery leg. A design that publishes in one account, forwards to another account’s bus, and then invokes a service has more billable path elements than a same-account publication delivered directly to a target.
| Event category | What to include in an estimate | Important qualification |
|---|---|---|
| AWS management | Downstream delivery, cross-account or cross-Region paths, and optional features | AWS currently describes ingestion as free; verify the live regional page |
| Partner | Ingestion plus every delivery leg and optional feature | Partner-event rates and source configuration determine the charge |
| Custom | Ingestion, delivery, forwarding, and optional feature usage | Payload chunks multiply event units |
Payload size multiplies event-bus units
AWS bills each 64 KB payload chunk as one event unit for event-bus pricing. Use the ceiling of payload size divided by 64 KB for each logical event. A payload of 1 KB through 64 KB is one unit; 64 KB plus even a small amount becomes two units. A 256 KB payload counts as four units.
Rank #2
For a practical estimate, group events by size rather than using only an average. For example, 8 million custom events at 20 KB each produce 8 million 64 KB units, while 2 million events at 256 KB each produce 8 million units despite having one quarter as many logical events. Include any serialization, envelope, or metadata that is part of the EventBridge event when determining the payload size.
Account and Region boundaries add route-specific charges
Delivery within one account is not necessarily priced like delivery to another account. Trace every hop: source account, source bus, forwarding rule, destination bus, and final service. AWS’s pricing examples show distinct treatment for cross-account delivery and for delivery to a service in the same account.
Cross-Region designs require two checks. First, standard AWS inter-Region data-transfer charges can apply to the bytes that move between Regions. Second, the EventBridge FAQ states a $1-per-million-event charge for replication of cross-Region events. Treat that replication figure as a specific pricing rule, not as a substitute for data-transfer pricing; confirm that your implementation uses the replication feature covered by the FAQ.
Keep account and Region boundaries in your architecture diagram and cost worksheet. A route that crosses both can incur delivery, replication, and data-transfer meters in addition to the original event-bus ingestion.
Rank #3
Archives and replay create storage and processing meters
An EventBridge archive can retain selected events for a configured period. Retention defaults to indefinite unless you set a period, so an archive that is enabled without a lifecycle decision can accumulate data continuously. AWS pricing lists storage, archive processing, and replayed-event charges as separate considerations.
Estimate the volume retained each month, the number of days or months it remains, and the amount you expect to replay. A replay is not merely a free read of existing data: replayed events return to the source event bus and can generate normal downstream delivery or target charges again.
AWS documentation notes several operational details that affect estimates and incident response:
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
- Events can take time to arrive in an archive after publication.
- AWS recommends waiting 10 minutes before replaying recently received events so they have time to reach the archive.
- Replayed events are not necessarily delivered in the order in which they originally arrived.
- Archive filters determine which events are retained, so filtering can materially change storage and later replay volume.
Use the archive and replay documentation alongside the pricing page when forecasting recovery or backfill workloads.
Rank #4
Schema Registry is free; Schema Discovery is metered
Do not combine these two features in a cost model. AWS states that using the Schema Registry is free. Schema Discovery, which ingests events to infer schemas, has a separate meter and a monthly free tier of 5 million ingested events. The FAQ lists an overage price of $0.10 per million ingested events.
Discovery uses 8 KB chunks, not the 64 KB chunks used for event-bus pricing. Calculate discovery units as the ceiling of payload size divided by 8 KB. A hypothetical 2 million events averaging 20 KB would produce 6 million discovery units (three 8 KB chunks per event); 5 million would fall within the stated monthly free tier and 1 million would be billed at the FAQ’s $0.10-per-million rate. This calculation is for discovery only and does not replace the event-bus calculation.
Schema Discovery’s free tier and price are volatile service terms. Recheck the current pricing page and FAQ for the Region and date of your estimate.
API Destinations can add invocation and network costs
API Destinations have their own invocation meter. AWS counts each 64 KB request payload chunk as a separate request, so a large outbound body can multiply invocation units just as a large event multiplies bus units.
Best Value
The invocation figure is not necessarily the complete cost. A public internet endpoint can incur standard AWS data-transfer charges. A private endpoint using AWS PrivateLink or VPC Lattice can incur those services’ connectivity and data-processing charges. Record whether the destination is public or private, the number of requests, request-size distribution, and the bytes transferred before comparing designs.
Pipes and Scheduler are separate products
EventBridge Pipes and EventBridge Scheduler appear on the same AWS pricing page but are not universal components of event-bus pricing. Pipes have their own request and payload-chunk meters. Scheduler has a separate invocation model; the pricing page currently displays an allowance of 14 million invocations per month in its free-tier information.
Include either product only when your architecture uses it, and model its requests or invocations separately from event-bus events. Confirm current rates and free-tier conditions on the live pricing page because they can change independently of bus pricing.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →How to build an EventBridge cost estimate
Collect these inputs before applying any rate:
- Classify every source. Mark each stream as AWS management, partner, or custom.
- Measure logical volume. Record events published and delivered during the billing period, including expected bursts.
- Bucket payload sizes. Use a distribution such as 0–64 KB, 64–128 KB, and so on instead of one optimistic average.
- Map destinations. List every rule target, bus-to-bus forward, and service invocation.
- Mark account boundaries. Identify each cross-account delivery and whether it ends at another bus or a service.
- Mark Region boundaries. Add inter-Region data transfer and determine whether cross-Region replication is enabled.
- Describe archive behavior. Specify filters, retained data per period, retention duration, and expected replay volume.
- Separate schema usage. Record whether Schema Discovery is enabled and the number and size of events it processes; leave the free Schema Registry out of the metered total.
- Describe API connectivity. For API Destinations, record request size, invocation count, public versus private endpoint, and any PrivateLink or VPC Lattice path.
- Include adjacent services deliberately. Add Pipes or Scheduler only if they are part of the deployed design.
- Set the Region and date. Use the relevant AWS Region and note that feature availability and rates can differ by Region.
Convert the inputs into billing units
- Event bus: for each payload, calculate
ceil(payload bytes ÷ 64 KB), then multiply by the number of events in that size bucket. - Schema Discovery: calculate
ceil(payload bytes ÷ 8 KB)per discovered event, total the units, subtract the 5-million monthly free tier, and apply the current overage rate to any remainder. - Delivery: count each applicable delivery leg and classify it as same-account, cross-account, or cross-Region according to the pricing rules.
- Archive: estimate retained gigabytes over time, archive processing, and replayed events rather than only initial publication.
- API Destination: count 64 KB request chunks and add connectivity or data-transfer meters for the chosen endpoint path.
Then run at least three scenarios: a normal month, a peak month, and a recovery or replay month. This exposes costs that a single steady-state average hides.
Architecture examples that change the result
| Architecture pattern | Cost items to model | Why a simple event count fails |
|---|---|---|
| Custom events to a service in the same account | Custom ingestion, 64 KB event units, target delivery, and any archive or discovery | Payload size and optional features still multiply usage |
| Partner events forwarded to another account’s bus | Partner ingestion, cross-account delivery, destination-bus processing, and downstream delivery | Origin and account boundary use different pricing paths |
| Cross-Region replicated bus | Source and destination paths, the stated replication meter, and inter-Region data transfer | Replication and network charges are distinct |
| Bus event sent to an API Destination | Bus units, API invocation chunks, and public or private network charges | One logical event can produce multiple request and network meters |
| Archive with periodic replay | Retention storage, archive processing, replayed events, and replay-triggered delivery | Historical data creates recurring storage and later processing |
Common estimation mistakes and fixes
- Counting logical events instead of chunks: calculate 64 KB bus units and, separately, 8 KB discovery units.
- Assuming management events make the whole workflow free: price delivery, network transfer, archives, and targets independently.
- Ignoring forwarding legs: draw the route through every bus and account before adding totals.
- Leaving archive retention indefinite by accident: set a retention policy and model the resulting storage curve.
- Budgeting replay as a one-time read: include replayed-event processing and downstream deliveries.
- Using the API invocation rate as the total API cost: add data-transfer, PrivateLink, or VPC Lattice charges where applicable.
- Adding Pipes or Scheduler to every bus estimate: include them only when deployed and use their separate meters.
- Copying a rate from another Region: check the applicable Region, date, and feature availability before approving the estimate.
Check regional availability before deployment
Pricing and feature availability are separate questions. A feature may be listed on the global pricing page while its availability differs by Region. Check the EventBridge feature-availability table, then confirm the Region-specific prices on AWS’s live pricing pages. Record the access date in your cost model so a later bill review can distinguish a rate change from a usage change.
A practical review checklist
Before you sign off an EventBridge estimate, verify that it contains:
- Event category for every source.
- Published, delivered, archived, replayed, and discovered volumes.
- Payload-size buckets and the correct 64 KB or 8 KB chunk rule.
- Every bus, target, account boundary, and Region boundary.
- Inter-Region transfer and any cross-Region replication charge.
- Archive filters, retention duration, storage growth, and replay assumptions.
- Schema Discovery usage separated from the free Schema Registry.
- API Destination invocation chunks and endpoint-network path.
- Pipes and Scheduler listed only when actually used.
- A Region-specific rate check dated October 13, 2026 or updated to the day you publish the estimate.
With those fields documented, you can explain not only what EventBridge costs, but which architecture decision caused each part of the bill.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
