Portworx Autopilot Release Notes
1.3.18
October 27, 2025
Improvements
| Issue Number | Issue Description |
|---|---|
| PWX-47044 | Added support for the dontWaitForCleanVolumes parameter in pool expansion actions to allow resize-drive operations even when volumes have only a single replica on the pool. For more information, see AutopilotRule CRD Reference. |
note
This release also addresses security vulnerabilities.
1.3.17
May 21, 2025
Notes
- This release addresses security vulnerabilities.
1.3.16
December 9, 2024
Improvements
| Issue Number | Issue Description |
|---|---|
| PWX-38984 | Added support for matchExpressions along with matchLabels in Autopilot rules. You can now select the objects using either matchLabels, matchExpressions, or both selectors. |
| PWX-34122 | Updated the CRD schema to adhere to the strict structural schema validations introduced in Kubernetes 1.23 and later. This allows setting actionCoolDownPeriod and pollInterval without encountering strict decoding errors. |
| PWX-38947 | Autopilot now supports add-drive and resize-drive options for pool expansion. Additionally, the default resize type is changed to auto. |
| PWX-35367 | Implemented a dedicated scheduler process for moving FSM states, which would decrease the latency of Autopilot operations. |
| PWX-39812 | The following changes has been made to reduce the CPU usage of the Prometheus pod:
|
Fixes
| Issue Number | Issue Description | Severity |
|---|---|---|
| PWX-39803 | If the rebalance action is skipped due to an error during the dry run without checking the audit actions, the cluster might be left in an unbalanced state. User Impact: Impacted overall performance and resource distribution across the cluster. Resolution: Added checks to ensure that rebalance is not skipped when audit entries are present in the dry run response. Affected Versions: All | Minor |
| PWX-36918 | Autopilot assumed the default .svc.cluster.local DNS domain for the Kubernetes cluster, causing connection failures using different DNS domains. User Impact: Connection failures were seen when using a DNS domain other than the default domain. Resolution: Autopilot now determines the DNS domain Kubernetes uses for its internal services. Affected Versions: All | Major |
| PWX-35742 | Autopilot triggered multiple dry run rebalance operations, even if there is a run in progress. User Impact: Multiple dry run rebalance operations were trigerred continuously. Resolution: Added checks to ensure that no other rebalance operation is running before triggering a new one. Affected Versions: All | Minor |
1.3.15
July 25, 2024
Notes
- In order to use Portworx with OpenShift version 4.16, you must use Operator version 24.1.1 and Autopilot version 1.3.15.
1.3.14
March 18, 2024
Improvements
| Improvement Number | Improvement Description |
|---|---|
| PWX-36274 | OpenShift users on version 4.12 or newer, can now configure Autopilot to use OpenShift Prometheus deployment for monitoring. |
Fixes
The following issues have been fixed:
| Issue Number | Issue Description |
|---|---|
| PWX-36160 | In certain scenarios, Autopilot would generate duplicate AutopilotRuleObjects for Storage pools. The pool expansion and re-balance rules would overlap, generating events on a common AutopilotRuleObjects. This overlap prevented any of the rules from proceeding and taking action.Resolution: Autopilot now does not create duplicate AutopilotRuleObjects for storage pools. |
| PWX-36046 | Prometheus self-signed CA certificate bundle was inadvertently overriding the container's default CA certificates. As a result, while Autopilot could integrate with SSL-enabled Prometheus instances using self-signed SSL certificates, it would fail to work with those using commercially signed SSL certificates. Resolution: Autopilot is now compatible with both commercial and self-signed SSL-enabled Prometheus endpoints. |
Known issues (Errata)
| Issue Number | Issue Description |
|---|---|
| PD-2814 | When attempting to resize a pool or volume using AutopilotRules, users may encounter errors leading to a failed condition check for a rule and see the connection reset by peer error message when in the Kubernetes events by running the command kubectl get events --field-selector involvedObject.kind=AutopilotRule --all-namespaces. Autopilot may enter a state where it does not retry the operation after encountering this error. Workaround: Manually restart the Autopilot pod within your Kubernetes environment. This action forces Autopilot to reconnect to the OCP Prometheus endpoints. |
1.3.13
February 08, 2024
Notes
- In order to use Portworx with OpenShift version 4.14, you must use Operator version 23.10.3 and Autopilot version 1.3.13. The Upgrade to OpenShift 4.14 or newer with Portworx page details the steps for upgrading OpenShift.
- Enabling monitoring on OpenShift version 4.14 will no longer setup a separate Prometheus instance for metrics and Autopilot.
1.3.12
November 30, 2023
Improvements
| Improvement Number | Improvement Description |
|---|---|
| PWX-34605 | Autopilot now appends the rule name as a suffix to newly created AutopilotRuleObjects. As a result, Autopilot can schedule and execute multiple rules for a pool without overriding them. |
| PWX-33620 | The time required to resize PVCs in a cluster with a large number of volumes has been considerably reduced. |
1.3.11
September 25, 2023
Improvements
| Improvement Number | Improvement Description |
|---|---|
| PWX-33668 | Autopilot now does not treat an offline Prometheus target as a fatal error. It continues to monitor metrics from other Prometheus targets. |
Fixes
The following issues have been fixed:
| Issue Number | Issue Description |
|---|---|
| PWX-33620 | With Autopilot, resizing of PVCs took a long time in a cluster with a large number of PVCs. User impact: Autopilot's Kubernetes API calls were being throttled adding to the delay required for triggering actions. Resolution: Added controller-runtime caching to Autopilot to reduce the time taken for volume resizing. |
1.3.10
August 16, 2023
Notes
- This version addresses security vulnerabilities.
1.3.9
August 7, 2023
Notes
- This version addresses security vulnerabilities.
1.3.8
May 31, 2023
Fixes
The following issues have been fixed:
| Issue Number | Issue Description |
|---|---|
| PWX-31468 | Autopilot pods kept restarting when deployed on OpenShift with the fsGroup securityContext setting.User impact: Autopilot pods crashed because it used a filesystem watcher to monitor configuration changes and load applications. Resolution: Autopilot now uses the sched-ops API to monitor configuration changes. |
1.3.7
February 22, 2023
Notes
- This version addresses security vulnerabilities.
1.3.6
February 16, 2023
Notes
- This version addresses security vulnerabilities.
1.3.5
January 26, 2023
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Number | Improvement Description |
|---|---|
| PWX-28744 | Miscellaneous improvements. |
1.3.3
September 26, 2022
Fixes
The following issues have been fixed:
| Issue Number | Issue Description |
|---|---|
| PWX-26135 | Parameters in AutopilotRule actions were pruned for Kubernetes versions 1.23 and higher. User impact: Users on Kubernetes 1.23 or greater saw some parameters, such as 'x-kubernetes-preserve-unknown-fields: true', removed when applying a CRD. Resolution: Autopilot no longer prunes parameters. |
1.3.2
December 14, 2021
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Number | Improvement Description |
|---|---|
| PD-1047 | Autopilot 1.3.2 now supports installations that use PX-Security. Follow the instructions in Autopilot Install and Setup to use Autopilot with PX-Security. |
1.3.1
November 22, 2021
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Number | Improvement Description |
|---|---|
| PD-1035 | Autopilot 1.3.1 now supports Kubernetes 1.22 and higher |
1.3.0
Sep 02, 2020
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Number | Improvement Description |
|---|---|
| AUT-113 | Added support to rebalance Portworx storage pools. Refer to documentation for more details. |
| AUT-136 | Added support to users to require approval before actions are taken for a AutopilotRule. Refer to documentation for more details. |
| AUT-138 | Added support for requiredMatches in the AutopilotRule CRD. This allows users to specify number of conditions that are required to match in the rule. See the Automatically rebalance Portworx storage pools use case on example on how this can be used. |
| AUT-144 | Added support for Autopilot to create Github Pull Requests to approvals for its actions. This enables users to integrate GitOps workflows with Autopilot. for more details. |
| AUT-157 | Make AutopilotRuleObject namespace scoped. Refer to documentation for usage of AutopilotRuleObjects. |
| AUT-205 | Improve support bundle collection |
Fixes
The following issues have been fixed:
| Issue Number | Issue Description |
|---|---|
| AUT-83 | Autopilot pool expand should never bring Portworx out of quorum User impact: In certain situations where 2 or more pools have non-intersection volumes, Autopilot can triggered expand on multiple pools at the same time which can bring the PX cluster out of quorum. Resolution: Autopilot will now perform expansion on only one pool at a given time. Subsequent pools will have their actions in pending state until the previous one is complete. |
| AUT-87 | If an action is declined, Autopilot will now perform exponential backoff before retrying it. User impact: If an action was declined due to the maxsize being hit in the AutopilotRule for a PVC resize, Autopilot used to aggressively retry flooding logs and events. Resolution: Autopilot will now perform exponential backoff. |
| AUT-169 | Autopilot loses track of volume for rule when PX-Backup backs up and restores to same namespace. User impact: If a user has a PVC that was restored from a backup or pre-provisioned, Autopilot would not be able to track the metrics of the PVC correctly. Resolution: Autopilot was incorrect assuming the volume name of a PVC by using the PVC UUID. Instead now, it will use the actual volume name from the PVC spec. |
| AUT-209 | Fixed an issue where Autopilot incorrectly determines that conditions are not met for an object. User impact: There are situations where the prometheus API reports empty metrics when it is initializing and it hasn't scrapped its targets. If Autopilot queries prometheus during such times, it will find empty metrics and will be misled to consider the conditions as not being met for that object. Resolution: Autopilot will no longer accept metrics from Prometheus if a Portworx metrics target is not healthy. The Portworx metrics target will become healthy when Prometheus has initialized completely and has been able to perform a successful scrape. |
1.2.1
May 30, 2020
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Description |
|---|
| The PVC resize action no longer requires a Portworx Autopilot Capacity Management license |
1.2.0
March 20, 2020
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Description |
|---|
Added support for a new parameter, scalesize, under the storage pool expand action that allows you to increase storage pool capacity by a fixed amount. See details on all action parameters here. |
| Added validation for the AutopilotRule CRD |
| Added an alert event that occurs when Autopilot detects an action that cannot be performed due to license restrictions. See the list of alerts for more information. |
1.1.0
February 19, 2020
Improvements
Portworx has upgraded or enhanced functionality in the following areas:
| Improvement Description |
|---|
| Added a new CRD, called AutopilotRuleObject, which can be used to check for useful events in objects that autopilot monitors, such as PVCs and StoragePools. |
| Added basic metrics for monitoring Autopilot and Grafana dashboards to view them. To view Autopilot metrics, follow the steps in the Prometheus and Grafana article. |
Fixes
The following issues have been fixed:
| Issue Description |
|---|
| When an Autopilot pod restarted while expanding a storage pool, it sometimes started expanding another storage pool. User impact: If multiple pools for a volume started expansion together, the volume could go out of quorum. Resolution: Autopilot pods now correctly wait for previous storage pool expansions to complete when they're restarted. |
| Deleting Autopilot rules or PVCs sometimes caused the Autopilot pod to crash User impact: If Autopilot crashed, it could start expanding another storage pool while one was already in being expanded. Resolution: The Autopilot pod no longer crashes when rules or PVCs are deleted. |
1.0.0
November 18, 2019
Introducing Portworx Autopilot! See the Autopilot section of the documentation for more information