Automation for Jira における低速なルール実行のトラブルシューティング
プラットフォームについて: Cloud と Data Center - この記事は クラウド プラットフォームとデータセンター プラットフォームの両方に等しく当てはまります。
Server* 製品のサポートは 2024 年 2 月 15 日に終了しました。Server 製品を実行している場合は、 アトラシアン Server サポート終了 のお知らせにアクセスして、移行オプションを確認してください。
*Fisheye および Crucible は除く
要約
Using the audit log, we might eventually find that some rules are taking a longer time to execute than expected.
The first step is to check the Optimize rules page to view general rule best practices.
Starting in Automation for Jira 7.4.0, we can export the log to JSON and find more details about the rule execution.
診断
ログのエクスポート後、JSON データをフォーマットして読みやすくします。Sublime や VS Code などのエディターにはそのためのツールがありますが、そのジョブを実行できるオンライン ページもいくつかあります。
分析を始めるために一例を挙げましょう。
{
"items": [
{
"id": 741068,
"summary": "Create issue on transition",
"authorKey": "jiraadmin",
"created": 1643673518072,
"startExecution": 1643673518072,
"endExecution": 1643673595145,
"category": "SUCCESS",
"eventSource": "jira.issue.event.trigger:transitioned",
"objectItem": {
"id": 1229,
"name": "Create issue on transition",
"typeName": "jira.issue.event.trigger:transitioned"
},
"globalAssociatedItems": [],
"globalMessages": [],
"componentChanges": [
{
"id": 1706137,
"componentId": 60056,
"component": "ACTION",
"componentNameKey": "Create issue",
"startTime": 1643673589202,
"duration": 11882,
"associatedItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 0,
"results": []
},
"changeItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 1,
"results": [
{
"fieldName": "Issues created successfully",
"changeTo": "SD-123"
}
]
}
},
{
"id": 1706136,
"componentId": 60054,
"component": "CONDITION_BLOCK",
"componentNameKey": "If block",
"startTime": 1643673539852,
"duration": 98700,
"associatedItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 0,
"results": []
},
"changeItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 1,
"results": [
{
"fieldName": "The following issues passed:",
"changeTo": "SD-122"
}
]
}
},
{
"id": 1706135,
"componentId": 60052,
"component": "TRIGGER",
"componentNameKey": "Issue event",
"startTime": 1643673518085,
"duration": 43534,
"associatedItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 1,
"results": [
{
"id": "154644",
"name": "SD-122",
"typeName": "ISSUE",
"parentId": "10001",
"parentName": "SD"
}
]
},
"changeItems": {
"limit": 9223372036854775807,
"offset": 0,
"total": 0,
"results": []
}
}
],
"projectIds": [
"10001"
],
"duration": 77071
},
{
"id":...Each one of the items is a rule execution.
データの先頭には、概要というよりは一般に UI で使用できるデータがあります。トラブルシューティングで最も重要なフィールドは次のとおりです。
summary - rule name
created - rule trigger time
startExecution - the execution start date
endExecution - the execution end date
category - the item status
All dates are in epoch times, in milliseconds. To convert to actual dates, use a command or an online tool such as https://www.epochconverter.com/.
To find the rule execution duration, just subtract the startExecution from the endExecution times.
The most interesting reason to export the log is to inspect details about each component, which are under componentChanges.
コンポーネントは下から上に並んでいるので、下部にトリガーが表示されます。
Each component has its duration in milliseconds and its startTime. This helps us understand which component is the main offender, or if there is a general slowness.
Under results, we can see the results from that component. This will tell us the actions effectively performed by each component and might help understand longer execution times.
ソリューション
すべての状況に対応できる標準ソリューションはないため、最も一般的なソリューションのいくつかを次に示します。
Check the Optimize rules page for rule best practices.
If using JQL in a component, try to use smart values. If not possible, check Factors contributing to JQL performance in Jira Server and Data Center.
If there's a considerable difference between the created and startExecution dates. This could mean that rules are being queued - that is, rule creation rate was higher than the server's capacity to process.
他のルールを最適化してみてください。
データベースのパフォーマンスを検証します。
If the server CPU is not high, consider increasing the number of threads for Automation for Jira in its configured Service Limits (
automation.processing.thread.pool.size.per.nodeproperty)ℹ️ To avoid general performance issues, perform this increase in small steps.
明確な理由もなく、すべてのアイテムに長時間かかっている場合
Configure audit log expiration, as it might have grown too large: Expire audit log items
データベースのパフォーマンスを確認します。
この内容はお役に立ちましたか?