fix(amazonq): fix connection manager deadlock when q ListAvailableProfiles is long #5547
Conversation
…files is long
`CodeWhispererStatusBarWidget` attempts to query active connections, but lock is being held while trying to list profiles
```
"AWT-EventQueue-0" prio=0 tid=0x0 nid=0x0 blocked
java.lang.Thread.State: BLOCKED
on software.aws.toolkits.jetbrains.core.credentials.DefaultToolkitConnectionManager@37995400 owned by "ApplicationImpl pooled thread 5 @coroutine#17104" Id=128
at software.aws.toolkits.jetbrains.core.credentials.DefaultToolkitConnectionManager.activeConnectionForFeature(DefaultToolkitConnectionManager.kt)
at software.aws.toolkits.jetbrains.services.amazonq.QUtilsKt.calculateIfIamIdentityCenterConnection(QUtils.kt:18)
at software.aws.toolkits.jetbrains.services.codewhisperer.customization.DefaultCodeWhispererModelConfigurator.activeCustomization(CodeWhispererModelConfigurator.kt:171)
at software.aws.toolkits.jetbrains.services.codewhisperer.status.CodeWhispererStatusBarWidget.getSelectedValue(CodeWhispererStatusBarWidget.kt:114)
at com.intellij.openapi.wm.impl.status.MultipleTextValues.beforeUpdate(IdeStatusBarImpl.kt:983)
```
```
- "coroutine#17104":BlockingCoroutine{Active}@7dced952, state: RUNNING [CoroutineId(17104), BlockingEventLoop@2da7461c]
at java.base/jdk.internal.misc.Unsafe.park(Native Method)
at java.base/java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:269)
at java.base/java.util.concurrent.CompletableFuture$Signaller.block(CompletableFuture.java:1866)
at java.base/java.util.concurrent.ForkJoinPool.unmanagedBlock(ForkJoinPool.java:4013)
at java.base/java.util.concurrent.ForkJoinPool.managedBlock(ForkJoinPool.java:3961)
at java.base/java.util.concurrent.CompletableFuture.timedGet(CompletableFuture.java:1939)
at java.base/java.util.concurrent.CompletableFuture.get(CompletableFuture.java:2095)
at migration.software.aws.toolkits.jetbrains.core.AwsResourceCache$Companion.wait(AwsResourceCache.kt:150)
at migration.software.aws.toolkits.jetbrains.core.AwsResourceCache$Companion.access$wait(AwsResourceCache.kt:145)
at migration.software.aws.toolkits.jetbrains.core.AwsResourceCache.getResourceNow(AwsResourceCache.kt:92)
at migration.software.aws.toolkits.jetbrains.core.AwsResourceCache.getResourceNow(AwsResourceCache.kt:105)
at software.aws.toolkits.jetbrains.services.amazonq.profile.QRegionProfileManager.listRegionProfiles(QRegionProfileManager.kt:70)
at software.aws.toolkits.jetbrains.services.amazonq.profile.QRegionProfileManager.validateProfile$lambda$0(QRegionProfileManager.kt:50)
at software.aws.toolkits.core.utils.ExceptionUtils.tryOrNull(ExceptionUtils.kt:11)
at software.aws.toolkits.jetbrains.services.amazonq.profile.QRegionProfileManager.validateProfile(QRegionProfileManager.kt:49)
at software.aws.toolkits.jetbrains.services.amazonq.toolwindow.AmazonQToolWindowFactory$createToolWindowContent$2.activeConnectionChanged(AmazonQToolWindowFactory.kt:68)
[...]
at software.aws.toolkits.jetbrains.core.credentials.DefaultToolkitConnectionManager.switchConnection(DefaultToolkitConnectionManager.kt:159)
```
| openMeetQPage(project) | ||
| QRegionProfileManager.getInstance().validateProfile(project) | ||
| ApplicationManager.getApplication().executeOnPooledThread { | ||
| QRegionProfileManager.getInstance().validateProfile(project) |
There was a problem hiding this comment.
i thoguht we already had this in project startup https://github.com/aws/aws-toolkit-jetbrains/blob/main/plugins/amazonq/chat/jetbrains-community/src/software/aws/toolkits/jetbrains/services/amazonq/startup/AmazonQStartupActivity.kt#L57
There was a problem hiding this comment.
do we really need this ?
There was a problem hiding this comment.
both --
so if we are checking in project startup, do we still need this one then?
There was a problem hiding this comment.
i feel we shouldn't need this call at all
There was a problem hiding this comment.
I think we might need to recheck the behavior when a user logs out and logs back in with completely different profiles. From my testing, validating only during project startup doesn't seem to be sufficient. Please feel free to correct me if I’m missing something!
There was a problem hiding this comment.
implement it in a different way
- log out clear the connection to profile arn cache, because anyway when you login, you have to reselect the profile
- remove
validateConnectiononConnectionChanged
There was a problem hiding this comment.
I see the reasoning behind clearing the cache on logout. That said, I think we still need to call listProfiles inside onConnectionChanged. Otherwise, prepareChatContent won’t be able to determine whether to open the WebView or Q panel. Let me know what you think.
https://github.com/aws/aws-toolkit-jetbrains/blob/main/plugins/amazonq/chat/jetbrains-community/src/software/aws/toolkits/jetbrains/services/amazonq/toolwindow/AmazonQToolWindowFactory.kt#L126-L127
There was a problem hiding this comment.
yes the list on prepareChatContent is separate from what this PR is addressing
Qodana Community for JVMIt seems all right 👌 No new problems were found according to the checks applied 💡 Qodana analysis was run in the pull request mode: only the changed files were checked Contact Qodana teamContact us at [email protected]
|
4 participants
CodeWhispererStatusBarWidgetattempts to query active connections, but lock is being held while trying to list profilesLicense
I confirm that my contribution is made under the terms of the Apache 2.0 license.