Using the Elasticsearch Add-in


Elasticsearch is a powerful open source search and analytics engine, and it’s easy to add to your application as an add-in.

We do not recommend changing the size of existing Elasticsearch servers as this can create critical issues and bring the server down. Follow our scaling guide instead.

Add Elasticearch

To add Elasticsearch to your application, access the add-ins menu, click Elasticsearch and choose the size of your new server.

This adds three environment variables to your application: ELASTICSEARCH_ADDRESS_INT, ELASTICSEARCH_ADDRESS_EXT and ELASTICSEARCH_ADDRESS, which you can use to connect to your Elasticsearch instance.

Upgrading Elasticsearch

Cloud 66 does not support in-place upgrades for Elasticsearch. Instead you should follow the checklist below to create a new Elasticsearch Cluster, and migrate your data to it.

Step 1: Create a new Elasticsearch server and group

You need to create a fresh set of servers running the latest version of Elasticsearch. To do this:

We will now build a brand new instance of Elasticsearch - you can watch the logs, or come back later (we will email you when the build is done).

If your instance of Elasticsearch is sharing your application server, you will need to clone your application as we cannot upgrade Elasticsearch in-place.

Step 2: Scale up more servers in the group (optional)

If your application requires more than one Elasticsearch server to handle the load, you should add them to your new group now. To do so:

We’ll now build another server for your cluster. You can repeat this step as many times as you require.

If you are running multiple Elasticsearch servers, for resilience it is recommended to run an ODD number of servers. Running an even number of servers increases the chance of data loss in the event of physical server failure.

Step 3: Migrate your Elasticsearch data to your new database group

Next, transfer your data from your existing Elasticsearch instance to your new group. We recommend following the official documentation to do so - be sure to select the correct version of your (existing) instance.

You can use Toolbelt (cx) commands like ssh , download and upload to make the process easier, or scp directly between your servers if you’re more comfortable with that.

Step 4: Promote your new group to “default”

Once your data is synchronised, you can switch your application to use your new cluster (group). To do so:

If your application relies on Elasticsearch to function, we recommend putting your application into maintenance mode before promoting your new cluster.

Step 5: Redeploy and test

Finally, redeploy your application to apply all the changes. Test your application is functioning as expected and then deactivate maintenance mode if everything is working. (If not, you can promote your old cluster to default and then fix any issues you’ve found.)