GA4 BigQuery Export Limit: 1 Million Events a Day and What to Do When You Hit It

If you have received an email saying your GA4 property has exceeded daily BigQuery export limits, this is the rule behind it. According to Google's Analytics Help, as of October 2026:

• Standard GA4 properties can send up to 1 million events a day through the daily (batch) export.
• Analytics 360 properties can send up to 20 billion events a day through the daily export.
• The streaming export has no event limit on either.

The limit counts events, not users or sessions. A site with modest traffic can still reach it if it fires many events per page: scroll tracking, video progress, custom interactions or a tag that sends everything twice. And it applies to the export, not to data collection: your GA4 reports carry on as before.

Google describes the sequence like this:

1. Property editors and administrators receive an email each time the property goes over the daily limit.
2. That email says when the export will be paused if nothing changes.
3. If the property keeps going over, the daily export is paused. If it goes over by a wide margin, Google says Analytics may pause it straight away.
4. Days that were not exported during the pause are not reprocessed later.

That last point is the one that hurts. The export does not fill in gaps: Google's help says data that has been exported cannot be re-exported, and that data from an export interrupted by an invalid payment method cannot be sent again either. A paused week is a hole in your BigQuery history for good, and it shows up months later in every year-on-year comparison. Why the export cannot go back in time.

Before excluding anything, look at the numbers. If the export is still running, this query counts events by name over one week. Replace my-project with your Google Cloud project and analytics_123456789 with your dataset, which is analytics_ followed by your property ID.

Events by name over a week

SELECT event_name, COUNT(*) AS events, COUNT(*) / 7 AS events_per_day

FROM `my-project.analytics_123456789.events_*`

WHERE _TABLE_SUFFIX BETWEEN '20261001' AND '20261007'

GROUP BY event_name

ORDER BY events DESC

The filter on _TABLE_SUFFIX reads only the seven daily tables, which keeps the query cheap, and leaves out the events_intraday_ tables, whose suffix starts with "intraday". If the export is already paused, the Events report in GA4 gives you the same ranking.

Some events hold up everything else. Excluding them saves volume but breaks the queries built on top:

• session_start and first_visit: they mark where sessions and new users begin, and many queries rely on them.
• page_view: most content and landing page analysis starts here.
• Your key events (formerly called conversions): they are usually the reason the export exists.
• user_engagement: it carries much of the engagement time data.

Good candidates are high-frequency custom events you do not analyse in BigQuery, scroll or video progress events if you only look at them in GA4 reports, and any test or staging stream.

The streaming export sends events to BigQuery within minutes, into a table called events_intraday_YYYYMMDD for the current day. It has no volume limit, so it is the usual way out for a standard property that has outgrown the daily export.

It comes with three catches, all from Google's own documentation:

• Cost. Streaming costs $0.05 per GB on top of normal storage and query costs. Google estimates that 1 GB is roughly 600,000 GA4 events, depending on event size.
• Completeness. It is a best-effort service with no completeness SLO, and it may contain gaps.
• Attribution. Traffic source data for new users and new sessions is not included, and for existing users the user-level attribution takes about 24 hours. Google recommends taking attribution from the daily export.

A rough calculation with Google's ratio: 3 million events a day is about 5 GB a day, around $7.50 a month in streaming charges. The money is rarely the problem; the missing attribution and the possible gaps are.

The BigQuery sandbox lets you use BigQuery without a credit card or a billing account, and Google allows the GA4 export to land there. Its limits, as stated in the BigQuery documentation in October 2026:

• 10 GiB of storage for the lifetime of the sandbox, which is not given back when you delete data.
• 1 TiB of queries a month, the same as the free tier.
• Tables, views and partitions expire automatically after 60 days.
• No streaming, no DML statements (INSERT, UPDATE, DELETE, MERGE) and no Data Transfer Service.

For GA4 this means no streaming export, and the daily tables disappear after 60 days. The sandbox is fine for learning the schema, but it does not lift the GA4 event limit, which belongs to the property, and it does not build a history. For that, the project needs billing enabled.

Linking GA4 to BigQuery costs nothing on a standard property. What you pay is BigQuery. These figures are from the Google Cloud pricing page in October 2026, in US dollars, for the EU and US multi-regions:

• Storage: $0.02 per GiB a month for active logical storage, and $0.01 for long-term storage, which applies to tables or partitions not modified for 90 consecutive days. The first 10 GiB each month are free.
• Queries: on demand, $6.25 per TiB processed, with the first 1 TiB each month free.
• Streaming: the $0.05 per GB described above, only if you turn it on.

Single regions can cost more: Madrid (europe-southwest1) lists $7.8125 per TiB. Choose the location with care, because changing it later means deleting the link and recreating the dataset.

For storage, at about 600,000 events per GB, a property exporting close to 1 million events a day adds around 50 GB a month: after a year, roughly $7 to $8 a month, since most of it is long-term storage. Queries are where bills grow. How to query the export without reading more than you need.