Once the load was appropriately balanced across our infrastructure, and we were able to confirm that our up-scaling systems were prepared to handle any future spikes, we re-enabled the Threads view for all users. The optimisation and redistribution work brought error rates down, which resolved most of the issues with channel loading, group direct messages, hyperlinks, and Saved items that users were experiencing.Īt this point, we made the call not to re-enable the Threads view in order to give our recovery efforts more time to restore balance to the system. Slack even added an extra popup which prevents leaving call without first closing a popup. We also made several code changes to reduce the effect a de-provisioning job of this scale would have on our infrastructure, and redistributed some of the load across other parts of the system. It is annoying to hunt down and do a multiple clicks to exit a call, when Zoom just does it automatically, or provides a one click button to do it all participants. In parallel, we ended the long running de-provisioning jobs, and significantly scaled up the systems that handle those jobs. ![]() ![]() Some customers may have also had trouble with loading channels, adding users to group direct messages, API hyperlink rendering, and missing Saved items.Ī mass de-provisioning event inadvertently increased load across our infrastructure, causing issues with Threads for many users.Īs an immediate mitigation strategy, we temporarily deactivated the Threads view for all users. ![]() PDT, a number of customers had trouble loading and receiving thread notifications.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |