app.js崩溃重启后，如何确保Jira问题实体属性实时更新？

Great question—this is a super common edge case for Jira Cloud add-ons that depend on webhooks to keep state in sync. Let’s break down practical, actionable ways to ensure your issue entity properties stay up-to-date after app.js crashes and restarts:

1. 启动时触发全量扫描与同步 #

As soon as your add-on finishes rebooting, kick off a full sync of issue link states:

• Call Jira’s REST API endpoint /rest/api/3/issueLink to fetch all existing issue links
• Iterate through each link, calculate the corresponding key-value pairs for entity properties, then batch-update the relevant issues using /rest/api/3/issue/{issueIdOrKey}/properties/{propertyKey}
• Don’t forget to set the "no links" state for issues that have no active links—this prevents gaps in your property data

2. Track Webhook Event Offsets for Incremental Catch-Up #

To avoid the performance hit of full scans (especially for large Jira instances), add event tracking:

• Use a persistent store (like a database, Redis, or even Jira’s own entity property storage) to record the timestamp or eventId of the last successfully processed webhook
• After restarting, pull all missed issuelink_created and issuelink_deleted events from Jira’s API that occurred between the last recorded timestamp and now
• Process these incremental events first, then run a full scan only if you suspect gaps (or skip it entirely if the incremental data covers all changes)

3. Reuse Your Existing connect_addon_enabled Logic #

You already have a working full-update flow for when the add-on is enabled—repurpose it for restarts:

• After reboot, manually trigger the internal function that handles the connect_addon_enabled event, instead of waiting for the actual webhook
• This lets you leverage tested, existing code instead of writing new full-sync logic from scratch

4. Add Periodic Validation & Repair #

Build a lightweight background task to catch any lingering inconsistencies:

• At regular intervals (e.g., daily), randomly sample a percentage of issues
• Compare their stored entity properties against their actual issue link state in Jira
• Fix any mismatched properties, and trigger a full sync automatically if the mismatch rate exceeds a threshold (like 5%)

Key Notes to Avoid Headaches #

• Respect Jira’s API rate limits—use batch requests and add retry logic with backoff to prevent being throttled
• Split full scans into chunks (e.g., 100 issues at a time) to avoid memory overflow
• Update your persistent event timestamp/ID immediately after successfully processing each webhook to minimize data loss

内容的提问来源于stack exchange，提问作者Priceline Negotiator