If your organization is migrating from one git host or ticket vendor to another, you must take steps in Flow to make sure your data remains accurate.
Depending on what data you’re migrating between vendors, follow the appropriate steps in Flow to maintain your data integrity.
The biggest risk when migrating data is Flow mistakenly duplicating your data. Use the steps in this article to avoid data duplication and make sure your data is showing up in the right places.
Migrating git hosts and only moving commit data from one vendor to another
If you’re only migrating commit data from one vendor to another, and not migrating PR or ticket data, you need to add a new integration for the new vendor.
Note: Do not make any changes to the original integration.
Flow uses a combination of the commit message and author information to identify commits, and has a process to deduplicate commits as necessary. As long as you’re moving commits from one host to another without any changes, Flow automatically removes duplicate commit data so your commits will not show up multiple times.
When only migrating commit data, you retain the historical PR data from your old integration, and get any new PR data through your new integration. Moving forward, commit data will come from the new integration, and old commit data will be deduplicated before being shown in reports, so it’s only represented once.
Links from Flow to your git host for old commits may not work if they’re pointing to your old vendor and it is no longer in service.
Tip: Flow has per-integration rate limits, not per-vendor, so adding a new integration will not cause additional rate-limiting concerns for your new integrations.
Migrating git hosts and moving both commit and PR data from one vendor to another
Flow does not deduplicate PRs, so if you’re migrating any PR data between vendors and add a new integration without modifying the old integration, you will have duplicate PR data in your metrics.
When creating a new integration, you can’t specify what timeframes to pull data from, so PR data duplication with multiple integrations is unavoidable.
To solve this, you must delete the original integration from Flow, then add the new integration. Flow will remove all data associated with the old integration, and show the historical commit and PR data through the new integration, as well as all new commit and PR data.
Note: When setting up a new integration, Flow only imports a set amount of historical data, depending on the vendor. Your new integration will only import data newer than the fixed data cutoff.
Removing the original integration and adding a new integration means your data will not be duplicated and your reports and metrics will be more accurate.
Important: If you only need to update credentials to an existing integration, do not delete the integration and add it back. Only delete the integration if you are moving from one vendor to a new vendor and need to create a new, separate integration.
The process of deleting data from the old integration and ingesting data from the new one is time-intensive. While it’s in-progress, some data might be duplicated and some data might be missing. It’s best to start this process before a weekend or during another time where you’re not actively using the data so you avoid using reports while data ingestion is in-progress.
Migrating ticket vendors
Flow does not deduplicate ticket data across vendors, so if you’re migrating any ticket data between vendors and add a new integration without modifying the old integration, you will have duplicate ticket data in your metrics.
When creating a new integration, you can’t specify what timeframes to pull data from, so ticket data duplication with multiple integrations is unavoidable.
To solve this, you must delete the original integration from Flow, then add the new integration. Flow will remove all data associated with the old integration, and show the historical commit and ticket data through the new integration, as well as all new commit and ticket data.
Note: When setting up a new integration, Flow only imports a set amount of historical data, depending on the vendor. You may not see data from as far back as you did with the original integration.
Removing the original integration and adding a new integration means your data will not be duplicated and your reports and metrics will be more accurate.
Important: If you only need to update credentials to an existing integration, do not delete the integration and add it back. Only delete the integration if you are moving from one vendor to a new vendor and need to create a new, separate integration.
The process of deleting data from the old integration and ingesting data from the new one is time-intensive. While it’s in-progress, some data might be duplicated and some data might be missing. It’s best to start this process before a weekend or during another time where you’re not actively using the data so you avoid using reports while data ingestion is in-progress.