Clarification on 'credit' for recent 20 OCT AWS outage
As we’ve had a few questions on the topic of credits, adding here for others/future reference:
Do Plugin People have ‘credits’ for platform outages as occurred 20 OCT
We don’t have any mechanism for such credits. Atlassian does refer its products now as ‘apps’ making this a little less obvious
Was it an app outage?
What was primarily affected was inbound routing to services hosted by AWS, like JEMHC, meaning the app was not accessible from Jira. This also mean that ‘webhooks’ containing ‘change’ data could not be delivered by Jira either, and were (we believe) lost (the delivery recovery mechanisms Atlassian has in place did not work for the same reason).
JEMHC ‘outbound’ connections were not affected by the problems so inbound mail retrieval continued without delay and we were able to process inbound mail normally (from mailhosts not running on AWS at least)
Could you please share your company’s Marketplace SLA policy regarding outages, including any criteria for credits or compensation?
We don’t have one currently, something we could to look at to show good will. Our hosting provider AWS, likewise doesn’t offer us compensation for outages, the last one would have been quite awkward.
We deploy our apps with high availability (multi node, multi availability zone) but a network/DNS issue such as the recent event breaks all that. We did start the DR process to another region but as the events unfolded, it became clear that actually triggering that would just result in more impact, so we decided to not fully swap over sites.