Frequently Asked Questions

Our pipeline will retain raw data for up to 12 hours in the event of emergencies that interrupt processing. This data lives only in a raw compressed queue and is not queryable/accessible easily.

Yes. Data is encrypted in transit and at rest. For Real-Time Metrics we store the aggregated results your filters produce, not your raw events: raw events sit in a compressed processing queue for at most 12 hours so we can recover from interruptions. AutoDocs stores event schemas and sampled property values to power the Events Explorer. Access is controlled with read/write-scoped API keys, per-user permissions and multi-factor authentication.

Some, yes, but ideally, you won’t notice. One of the motivating factors behind Aggregations.io was to remove any burden of thinking about cardinality. Read more about cardinality here

Almost definitely yes! If you’re using Segment or a similar CDP, they probably have a webhook destination. If you’re using Kafka, you can set up an HTTP Sink. As long as you can POST us JSON, we can probably ingest it. If you have a tricky situation, reach out and we can probably figure it out.

We’re a new company, it makes sense to be wary. We have worked in Data, Operations, Analytics, Product & more at some of the largest companies in the world. We have built Aggregations.io based on that experience, and we’re trying to provide products similar to tooling those companies have internally - for everyone.

If you need to ingest a substantial amount of events or have particular requirements, we may be able to work something out. Get in touch.

You can enable overages for your account, which will grant you an additional 10% of your plan’s event count each time you exceed the threshold, and are billed at a 10% premium of your subscription rate. This charge will appear on your following month’s invoice.

Absolutely! Admins of the organization will receive an email when your usage surpasses 50%, 75% & 90% of your event count.

A filter is the building block for your real-time aggregation. It defines what data matters, how you want to group your aggregations and at what interval. Learn more here

You can add up to 10 aggregations per filter. Each aggregation can contain 1-6 calculations, so you don’t need to worry about separate aggregations for Counts vs Sums etc.

With the Metrics API you build each request yourself: which filter, aggregation, calculation, time range and groupings. An Endpoint saves those choices in the UI, so callers only POST to its URL. You can change an Endpoint’s timing, fields or sorting without touching the code that calls it.

Send an x-api-token header with an API key that has Read permission. API keys can read your organization’s data, so call Endpoints from your backend rather than directly from a browser or mobile app.

Yes. Add a grouping for your customer (or account, team, etc.) on your filter and set it to Query Time on the Endpoint. Your backend then passes that customer’s ID in groupingFilters, and only their results come back.

JSON by default, or CSV by sending "format": "CSV". CSV responses include a header row, so they drop straight into spreadsheets and BI tools.

Yes. Pass an end timestamp in your request and the Endpoint’s duration is measured back from there. You can also set a fixed offset on the Endpoint, or on individual return fields to compare periods side by side, like today vs. yesterday.

Lookback Endpoints return up to 1,000 items by default. Enable sorting to control the order and set your own limit, including a limit on inner groupings, such as the top 5 players per team.

Safe changes go through. If a change would remove the last valid return field from an Endpoint, it’s blocked. Removed groupings come back empty and removed calculations return null, and the Endpoint editor highlights anything you need to fix. See the docs for details.

You can define an event by providing up to 3 properties that go into an event definition. Read more in the docs

Yes! Security and privacy is of the upmost importance. Review the docs on how to do this.

There are. In order to safegaurd the system (for example, if you define a user_id field as an event descriptor) - any config that has more than 250 unique events or 200 unique versions will be disabled. These are flexible limits, if you have a need for more events/versions, please contact us.

By default, events will be explored up to 10 levels deep, including nested arrays. This is a flexible limit, please contact us if you need this limit increased.