Source: Exporting Crashlytics data to BigQuery from Firebase
Crashlytics currently processes more than 3 billion events per day for our developers and distills them into an opinionated and actionable set of issues in our dashboard. We’ve heard thousands of requests over the years to let you explore further into your crash data, and that’s why today we’re extremely excited to offer the ability to export your Firebase Crashlytics data into BigQuery for deeper analysis!
With just a few clicks from the Firebase Crashlytics dashboard you can enable daily exports of all raw crash data on a per-app or per-project basis. This includes your stack traces, logs, keys and any other crash data.
The Crashlytics dashboard currently retains data for 90 days. With BigQuery you own the retention and deletion policies, making it much simpler for your team to track year-over-year trends in stability data.
BigQuery also offers the ability to export your data in CSV, JSON, or Avro format. This should help to streamline any GDPR data takeout requests you may receive.
Up until now, exploring your crash reports by custom metadata like Experiment ID or an Analytics breadcrumb has been limited, making it tough to identify which variant in an experiment is least stable or which level in a game has the most crashes. Now, when you export your data to BigQuery, it’s easy to run any deep analysis you want, and then visualize your report with Data Studio or any other business system you use.
As an example, say that you set up your Android game so that you log what level a crash occurs with:
When you export your Crashlytics data to BigQuery, you can then see the distribution of crashes by level with this query:
COUNT(DISTINCT event_id) AS num_of_crashes,
key = "current_level"
If you are a current Fabric user, you can gain access to BigQuery export and all the other features of Firebase by linking your app in the Fabric dashboard. Check out this link for details and documentation.
We hope this improvement makes it even easier for you to dig into your crash reporting data and efficiently debug your app! As always, if you have any questions, you can find us on Twitter and on Stack Overflow. Happy debugging!