Upgrade Universal Profiling
editUpgrade Universal Profiling
editThis functionality is in beta and is subject to change. The design and code is less mature than official GA features and is being provided as-is with no warranties. Beta features are not subject to the support SLA of official GA features.
Universal Profiling is a beta program, and we are developing and iterating over it frequently upon receiving feedback. For this reason, you have to erase all existing profiling data from your deployment when upgrading between minor releases in the Elastic Stack. When the product hits general availability, we will remove this requirement.
Upgrade process
editTo upgrade Universal Profiling, you need to:
- Perform a stack upgrade in Elastic Cloud
- Stop profiling data ingestion
- Delete existing profiling data
- Setup Universal Profiling from scratch
- Start profiling data ingestion
Perform a stack upgrade in the Elastic Cloud console
editTo perform a stack upgrade in Elastic Cloud:
- Locate the deployment you wish to upgrade in Elastic Cloud.
- Click the Manage deployment icon next to your deployment to open the deployment settings.
- Find Deployment version on the right side of the page.
- Click Upgrade, and choose the desired version.
- Verify the upgrade was successful under Your deployment → Activity in the navigation menu.
Stop profiling data ingestion
editDuring the Universal Profiling upgrade, you must stop data ingestion.
If you have an automated host-agent setup and control of the full fleet of machines where the host-agent is installed, deleting all host-agent deployments stops ingestion.
If you do not have direct control of all host-agent deployments, you can stop data ingestion by stopping incoming requests to the Integrations Server. To do this:
- From the deployment settings, scroll down to the Instances section and locate all Integrations Server instances.
- Click the three vertical dots in the upper-right corner of the Integrations Server card.
- Select Stop routing requests.
- Repeat this process for all Integrations Server instances.
When stopping incoming requests, host-agent replicas back off and retry connecting to the Elastic Cloud endpoint at increasing time intervals.
Delete existing profiling data
editYou can delete existing profiling data in Kibana:
- From the navigation menu, go to Stack Management → Index Management.
-
Make sure you’re in the Indices tab, and search for
profiling-in the search bar. - Select all resulting indices, click the Manage indices button, and select Delete indices* from the drop-down menu.
-
Switch to the Data Streams tab, and search for
profiling-in the search bar. - Select all resulting data streams, and click the Delete data streams button.
Verify that no ingestion is happening by reloading the Indices and Data Streams pages and ensuring that there are no indices or data streams with the profiling- prefix.
Set up Universal Profiling from scratch
editRefer to Get started to set up Universal Profiling from scratch.
Start profiling data ingestion
editIf you stopped ingesting data by stopping routing requests to the Integrations Server, reenable traffic following the same steps but click Start routing requests instead of Stop routing requests.
Verify the update succeeds
editClick any subheadings under Universal Profiling in the navigation menu. You should see incoming data.
If you see instructions on how to deploy the host-agent like in the examples from the Get Started documentation, the agents did not reconnect to the Integrations Server replicas.
Refer to the troubleshooting documentation and the Get Started documentation to investigate the issue.