Add exponential backoff to database reconnects

[kriswehner]

Right now, when the server(s) lose their database connections, they retry their reconnects as fast as possible.

What we want instead is an exponential backoff up to a maximum of around 5-10 seconds.

[gbrail]

Do you mean the snapshot or change server? The change server should already
do exponential backoff of the PG connection. Unfortunately we don't have an
automated test for this but I have tested it manually.

On Mon, Nov 21, 2016 at 10:11 AM, Kris Wehner [email protected]
wrote:

Right now, when the server(s) lose their database connections, they retry
their reconnects as fast as possible.

What we want instead is an exponential backoff up to a maximum of around
5-10 seconds.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#30, or mute the thread
https://github.com/notifications/unsubscribe-auth/AAf0a3qRlTHCJSqTwUHUDH72eHOwTrtTks5rAd7MgaJpZM4K4hWM
.