Updated March 2023
There are plenty of posts and documentation on the nitty gritty approaches for migrating to OpenSearch (i.e. rolling updates, snapshots, etc.). Start with the OpenSearch documentation for that.
Here we are covering something much larger and more important: The plans and support that need to be in place to facilitate a successful migration to OpenSearch.
What we are covering here will hold true whether you are migrating from Elasticsearch to OpenSearch, migrating from Splunk to OpenSearch, migrating from Oracle to OpenSearch, or from any other database.
Why migrate to OpenSearch?
OpenSearch is free to use and is fully open source under the Apache License, Version 2.0 (ALv2). These are the two main drivers for moving to OpenSearch.
Your organization can save a lot of money by not paying a licensing fee for Splunk, Oracle, Elasticsearch, or another paid database.
Keep in mind though that some new costs will arise, albeit likely totalling around 5-10% of the total savings. You’ll end up likely saving 90% or more of your paid license fees, but sometimes companies are surprised that some investment is required. We’ll get more into the investment below.
Do you have a team in place to support the migration?
When migrating to OpenSearch there is going to be a lot that is new and unknown. For instance, all security related features are configured differently in OpenSearch than Elasticsearch.
We’ve worked with many teams on OpenSearch migration, and we’ve seen that even the most stellar, well-read, self-learners need guidance and training on OpenSearch. It’s a relatively new tool with limited documentation, and at this time it’s not possible for a team to flawlessly migrate and optimize OpenSearch without expert help.
This support can come in a number of ways – hiring a new team member that is experienced with OpenSearch, enlisting outside consulting from OpenSearch experts, or signing up for managed OpenSearch where the migration is included as part of the project.
For more information on Dattell’s consulting and managed OpenSearch click here.
What is your plan for troubleshooting an outage?
Firstly, ensure you are backed up before starting a migration.
Secondly, set up monitoring before the migration to aid with optimization and troubleshooting.
Thirdly, continue running the original database for a few days to allow for a quick rollback if something unexpected occurs.
What are your tolerances for downtime?
Each use case has its own specific requirements. Opensearch has many features that increase uptime at the cost of additional hardware usage.
Is there clear documentation for all of the configurations you’ll need?
The OpenSearch documentation is getting better, but it’s still nascent. For instance, at the writing of this article there is limited or unclear documentation for setting up OpenID authentication. Getting the support of a person or team who has been through the migration and optimization before will help this process move much more smoothly.
Do you have a plan for re-optimizing your clusters once they are migrated to OpenSearch?
You might find that your engineers who were flawlessly optimizing and managing the old database are now struggling with doing the same with OpenSearch. It typically comes down to two issues.
Firstly, OpenSearch is its own unique database. If you’re coming from Splunk or Oracle it’s going to be very different. If you’re migrating from Elasticsearch it will be more similar but still have changes.
Secondly, because of the limited documentation and community support mentioned above, your team will be limited in its ability to self-learn and research solutions. Without the right support (see above), there can be a lengthy trial and error period that leaves both engineers and management frustrated.
With the right support in place, your team can get up-to-speed quickly with how to use, optimize, and troubleshoot OpenSearch.
Is it better for your use case to move incrementally or all at once?
Whenever possible, we recommend moving projects or teams incrementally rather than all at once.
Who is training end users and developers?
How is OpenSearch different from Splunk?
Opensearch supports nearly all Splunk features, just in a different way. The graphical interface, storage, data types, and user interface are all different.
How is OpenSearch different from Elasticsearch?
We have a detailed post on OpenSearch vs. Elasticsearch that breaks down their similarities and differences.
In brief, if you are moving from free Elasticsearch, then you’re going to be getting some extra features that are included in OpenSearch but only available under the paid versions of Elastic.
If you are moving from X-Pack / licensed Elasticsearch, some features might not be baked into OpenSearch. In that case, some additional work may be required.
Have OpenSearch Questions?
Managed OpenSearch on your environment with
24/ 7 support.
Consulting support to implement, troubleshoot, and optimize OpenSearch.
Published by