Starting next week, we will be upgrading our customers to Elasticsearch 1.3. This upgrade brings a slew of improvements to Elasticsearch, a new scripting engine, and some important improvements to our own underlying architecture.
- Monday, 11 August 2014 — Starter and Staging clusters.
- Monday, 18 August 2014 — Production clusters and up.
We waited to upgrade to Elasticsearch 1.3 directly, skipping 1.1 and 1.2, in part due to the changes made to dynamic scripting support. Notably the switch to using Groovy as the default scripting engine. The Elasticsearch blog has more details about the changes and improvements to scripting in their recent post, All About Scripting. Customers making use of dynamic scripting should take particular care to verify application compatibility when Starter and Staging clusters are updated next week.
While the upgrade to 1.3 should be backward-compatible for the majority of use cases, it's always a good idea to test your applications. We recommend that our customers update Elasticsearch in their development environments, as well as test against a Starter or Staging cluster next week to verify their application's compatibility. Particularly if your app uses dynamic scripting.
Any customers with concerns about compatibility may email us at firstname.lastname@example.org for an extension on their migration. We also have an option to host clusters on a single-tenant architecture for anyone who needs long-term version support.
In addition to the many new features and improvements to performance and resiliency in Elasticsearch itself, this upcoming migration include some major improvements in our own underlying architecture. We'll write about these improvements in more detail starting next week as they start becoming accessible.