Our uptime in December was greater than 99.999% after accounting for a very slight increase of 503 errors for the month on a very small number of tenants. We expect an upcoming rollout of Elasticsearch 1.4 will help us eliminate these errors, and until then we are capping our uptime at five nines for the month.
During December we did launch our first Elasticsearch 1.4 clusters into our AWS US East region. All new clusters in that region have been provisioned on Elasticsearch 1.4 to overall positive reception.
Development for Elasticsearch 1.4 had focused specifically on resiliency, and so brings a number of performance and availability improvements. We’ll be announcing plans for broader upgrades to Elasticsearch 1.4 within the next few weeks.
For further information on how we think about and report our availability, see our article on Measuring Availability as an Elasticsearch Hosting Provider.