Elastic AutoOps now free for all: What you get with it

Simplify your Elasticsearch operations with real-time issue detection and actionable recommendations to optimize performance and reduce costs. AutoOps is available for cloud and self-managed deployments. Learn more about AutoOps.

Same AutoOps, regardless of your license

We’re making this simple: Starting today, AutoOps is available at no cost for every self-managed Elasticsearch cluster through Elastic Cloud Connect. Whether you’re on Free, Basic, Platinum, or Enterprise, you get the same feature product. This isn’t a limited preview or a "lite" version. It’s the same product used by the largest deployments.

An investment in the self-managed community

Extending it to every user, including those on the free distribution, reflects Elastic's commitment to the success of the entire Elasticsearch community. By providing AutoOps for free, we’re investing in the stability and performance of the hundred of thousands of clusters that power the community's search and analytics workloads.

Elastic Cloud Connect enables self-managed clusters to consume Elastic Cloud services, such as AutoOps and the recently announced Elastic Inference Service (EIS), without the operational overhead of maintaining, patching, monitoring, and operating the services locally.

How AutoOps for self-managed works and what it provides

As clusters grow in complexity and size, you find yourself spending more time chasing configuration tweaks and trying to find the root cause of an issue. Monitoring tools show you metrics and leave the manual correlation to you and your favorite large language model (LLM) to find the root cause when the issue appears. AutoOps tells you what’s wrong, why, and exactly how to fix it, with real-time issue detection and specific resolution paths.

AutoOps runs on Elastic Cloud; there’s no infrastructure for you to provision or maintain. You simply run a lightweight agent on-premises to connect your cluster to the AutoOps service where operational metadata (such as node stats, cluster settings, and shard states) is shipped to AutoOps in real time to provide insights and recommendations. Your data never leaves your environment.

AutoOps vs. Stack Monitoring for self-managed users

Stack Monitoring provides the essential telemetry and basic monitoring for your nodes and indices, showing you the trend over time for various metrics, and alerts you when thresholds are crossed, yet it often leaves the diagnostic burden on the engineer. AutoOps offers a complete picture of cluster health by correlating all relevant metrics. This provides valuable insights and clear instructions on how to resolve issues when they occur.

Faster root cause analysis

Your cluster was humming along, but it suddenly got red in the middle of the night (and, as usual, nothing changed the days before).

• With Stack Monitoring: A built-in alert will notify you when your cluster health turns red. To find the cause, you need to dig into your cluster logs and turn to Dev Tools to look at your shard allocations to understand why that primary shard couldn’t be allocated. Looking at your alerts history, you see another one informing you that your hot nodes reached 80% disk utilization two days ago. You can’t find any disk usage charts to learn more about your disk fill rate, you only know that your disk reached 90% utilization in the meantime, and when one of your data streams needed to roll over, a new backing index was created, but no shards could be allocated to any of your nodes.
• With AutoOps: The system notifies you when your cluster turns red (1). Looking at the timeline, you immediately realize that this happened because you failed to take action on the previous watermark events that AutoOps raised, namely the high watermark events (2) that started happening recently, and the low watermark ones (3) that built up over the past few days. It is now straightforward for you to know what you need to do to get your cluster back to green.

Higher signal to noise ratio

Keeping your Elasticsearch cluster healthy is most probably your main concern. Yet, it’s not uncommon for the health status to sometimes flap between green and yellow (and sometimes red), and the cause is not always worthy of your time.

• With Stack Monitoring: The built-in “Cluster Health” alert will continue to be raised on each health transition from green to either yellow or red. In some situations, like frequent index creations, this can create a lot of repeated and undesired noise. Also, and more importantly, there’s no distinction between a yellow and a red status.
• With AutoOps: There are dedicated “Status Red” and “Status Yellow” events, with different severities. The latter can be customized in many different ways to fit your use case, as shown in the screenshot below:
  1. Since the cluster health can turn yellow only for a brief duration, you can decide for how long to ignore the yellow status before being notified (for example, five minutes, in the screenshot below).
  2. Furthermore, there are a lot of legitimate operations that Elasticsearch does all the time and that make the cluster turn yellow. You can pick any of the operations you don’t want to be notified about when your cluster turns yellow because of them (for example, adding replicas, relocating replicas, or closing or opening an index, among others).
  3. Finally, and most importantly, if you have several clusters to manage, you don’t need to configure this for all of them separately; you simply decide to which ones this configuration should be applied. Simple, powerful!

More insightful correlations and comparisons of node metrics

When your cluster runs on more than just a handful nodes, you often need to see how they perform against each other, especially when looking at search and indexing performance.

• With Stack Monitoring: The Nodes list doesn’t allow you to focus on a specific data tier and doesn’t provide any search or indexing performance metrics that you can sort on. These metrics are available, but only once you drill into a specific node, which doesn’t allow you to easily compare node performance against each other.
• With AutoOps: The Nodes view allows you to select nodes from a specific data tier. It also provides you with over 50 metrics visualizations, among them search and indexing performance, which give you exactly the visual cues that you need to understand how each node performs against each other and whether there are any struggling nodes that require your attention. In the screenshot below, we can see that some nodes are indexing at double the rate of some others and that search latency is building up at four times the latency of the fastest nodes.

Quick overview of the main differences

Here’s a glimpse at the notable differences between AutoOps and Stack Monitoring, but you can find a more detailed breakdown in our official documentation.

Start now: Five-minute installation

Connecting your cluster takes minutes, regardless of your license type:

1. Log in to your free Elastic Cloud account, or sign up for one.
2. Choose how to connect your cluster: Elastic Cloud on Kubernetes (ECK), Kubernetes, Docker, or Linux.
3. Enter your Elasticsearch cluster endpoint, and run the single command to install and run the lightweight Elastic agent.
4. Access AutoOps in your Elastic Cloud account.

For more details on AutoOps and instructions on connecting your self-managed cluster, read our product documentation.

Reach out if you have any questions

Feel free to reach out to us to share your questions and ideas via our Slack community, by posting on our Discuss forum, or by clicking the “Give Feedback” button on the AutoOps product page. If you’re connecting a paid self-managed Platinum or Enterprise cluster, you can contact support within your Elastic Cloud account.

Read more

If you’re interested to learn more about AutoOps and what it can do for you, please head to the official AutoOps documentation and the following Elastic Search Labs articles:

• AutoOps makes every Elasticsearch deployment simple(r) to manage
• AutoOps: A journey to simplify self-managed Elasticsearch management
• Leveraging AutoOps to detect long-running search queries

Start using AutoOps for free