Frequently asked Flow on-premises questions

Tags: Flow on-prem

What kind of TLS/SSL certificates should I use?

They type of certificate you use depends on your company policies and procedures.  You should read over the Obtaining Certificates document for more information about how to obtain these certificates.  It provides valuable information about the differences.

Why does Flow Enterprise require an e-mail server?

We require an e-mail server so that we may provide functionality such as:

  • Inviting new users to the system.
  • Securely resetting user passwords.
  • Sending important system e-mails

Without this functionality, Flow Enterprise would not function well or provide you (or your users) with a good user experience.

Why are the GitHub, GitLab, and BitBucket cloud offerings configured differently than their enterprise counterparts?

 GitHub, GitLab, and BitBucket cloud offering are configured differently in Flow Enterprise because of how we communicate with them.  Due to this, we need the initial configuration for these offering running when the system starts.  For the enterprise/on-premises offerings, we can make this configuration much simpler for our users and contain them in the application.

Why do Flow Enterprise and Replicated processes run as root?

Flow and Replicated run as "root" inside secure docker containers.  

Why don't I see options to configure GitHub Enterprise or BitBucket Server during setup?

This is a common question.  The simple answer is, the setup is different than the configuration for the cloud versions of these products.  

Does Flow leave any TCP connections open, or does it disconnect immediately from the server, even if the server is down or some other error occurs?

Like many TCP clients, Flow will attempt to reuse a connection during a single data ingestion cycle.  However, that connection can only be reused/left open if the connection is successfully established.  If the server is down, the connection cannot be reused.

If the Git provider server is down, does Flow keep trying to connect ad nauseam or does it try back later? If it tries back later, for what time interval does it wait before it tries again?

We will attempt to connect at most once every other hour, per repo and per integration.  If the server is down, we will wait and retry at the next cycle.  If there are too many failures in a row, the connection is marked in an error state and not used again until manual intervention.

Might the connection of Flow cause any CPU or memory spikes on the Git provider server?

Possibly, but not very likely.  We will inquire about the list of available repos.  We will inquire about updates to pull requests.  We will perform a git clone or git pull.  Spread over 2 hours, these operations are usually not enough to impact the server.

How long does it take for PR's to show up?  How do WebHooks impact that?

Without Web Hooks:  In a healthy system without any queue backups, the PR will be visible within approximately 4 hours and the metrics will be fully calculated within approximately 5 hours.   

With Web Hooks:  In a healthy system without any queue backups, the PR will be visible within approximately 10 minutes and the metrics will be fully calculated within approximately 90 minutes.

Note: Numbers presented are for  2020.2.2 and above and reflect timings from our SAAS offering.  GitLab and Stash users will see additional delays in full metrics processing due to API differences in the those products.

back to top


If you need help, please email support@pluralsight.com for 24/7 assistance.