On Saturday, June 4th starting at 11:00pm PDT (June 5, 06:00 UTC) and lasting for up to four hours we will be performing significant upgrade work on our core routers in our primary datacenter near Seattle, WA.

Due to the redundant nature of our core routing infrastructure, throughout this window our network will remain up and running and accessible to the internet.  However there will be four occasions throughout this window (as each router is taken out of production, upgraded, and re-added to production in sequence) where some customers may experience sub-optimal routing and minor packet loss as the various routing protocols re-converge on different paths over which to send the network traffic.  These should be limited to 20-45 seconds apiece.

The purpose of this upgrade is to stay ahead of the curve as far as network capacity is concerned, as well as to keep up with the latest networking technologies.

This major maintenance event should be finished no later than 3:00am PDT (10:00 UTC) on June 5th.

Update, 4:15am PDT (11:15 UTC): Tonight’s major maintenance is complete.  We’re happy to report that the physical upgrades were completed successfully, with only 3 windows where sub-optimal routing and some packet loss occurred.  During this window, we did not have time to complete the transition of our IPv6 connectivity to our core, so we will be completing that work tonight starting at 11:00pm PDT (06:00 UTC, June 6th).  During this additional maintenance window, our IPv6 connectivity will flap a few times, however, none of our IPv4 connectivity or physical layer connections should be affected by this work.