FAQ & Troubleshooting
Frequently asked questions about Statsig data warehouse ingestion, including supported warehouses, scheduling, costs, and troubleshooting connection issues.
What IP addresses does Statsig access data warehouses from?
Statsig accesses data warehouses from both the Statsig console service and Statsig data pipelines. If your data warehouse is IP protected, refer to the Statsig IP range documentation for IPs to allowlist. Contact Statsig on Slack if you have any issues.Does event data from ingestion count towards User Accounting Metrics?
No, event data from ingestions doesn't count towards Statsig's User Accounting Metrics such as DAU or Retention. Customers typically send Statsig a subset of their events, which could result in multiple competing values for fact data such as daily active users in your Statsig project. Statsig recommends sending your own precomputed metric for DAU or as a daily event per user (1 daily_active event if a user was active that day).
How long does data take to load?
For most customers, data ingestions take 1-2 hours to materialize in the Statsig console after you schedule the ingestion. The schedule is in PST, not PDT, so ingestions may start an hour later or earlier depending on daylight saving time.
Does Statsig load data incrementally every day?
Statsig loads data incrementally every day. Statsig also monitors data over several follow-up windows for up to two weeks, and reloads data for a given day if the data has changed more than 1%.
Can I ingest multiple metrics in the same scheduled ingestion?
Yes, you can ingest multiple metrics (and event types) in the same scheduled ingestion. Statsig enables you to run a SQL query against your data warehouse cluster to join multiple tables and generate a view with all your precomputed metrics. Use this view as the source for your scheduled data ingestion to import multiple metrics at the same time.
For example, your dataset could import both metric-1 and metric-2, with metric-2 including multiple units of analysis, such as user_id and alphabet_id.
How do missing metric values affect experiment calculations?
If the metric value is unavailable for a given user on a given day, Statsig treats it as zero for additive metrics such as counts and sums. For metrics that depend on a user participating in the metric, such as conversion rate, Statsig excludes the user. Additive metrics typically have a single metric_value column in the ingested data, while ratio (participating) metrics typically have separate numerator and denominator columns.
Does Statsig notify about ingestion status?
Statsig shows the status of your daily ingestion on the console under the Ingestions tab. Statsig reports three kinds of ingestion statuses:
- ingestion succeeded for a given day
- ingestion succeeded for a given day, but no data was detected
- ingestion failed for a given day
Does Statsig automatically backfill data?
Statsig looks back 3 days from the initial ingestion to check whether data has changed (>5% increase in the number of rows) and automatically triggers a backfill if it has. Outside of this window, the customer must trigger a backfill for the range of dates.
Was this helpful?