vbhanuchander-lang commented on PR #9042:
URL: https://github.com/apache/devlake/pull/9042#issuecomment-5274196582

   Thanks for confirming it against real data — `created_at` stuck at 
first-insert while
   `updated_at` moves with re-extraction is exactly the signature, and it's 
useful to know it
   scaled with board size for you.
   
   One thing worth planning for before you validate: **upgrading alone will not 
backfill the
   rows you have already lost.** The new filter is still only applied on 
incremental runs:
   
   ```go
   if stateManager.IsIncremental() {
       since := stateManager.GetSince()
       if since != nil {
           clauses = append(clauses, dal.Where("..._changelog_items.updated_at 
>= ? ", since))
       }
   }
   ```
   
   Your historical gaps have an `updated_at` older than the `since` of the next 
incremental
   run, so they stay outside the window. Two ways to recover them, either is 
fine:
   
   - Run a **full-refresh sync** on the affected boards. `IsIncremental()` is 
false, no time
     filter is added, and every collected changelog item in board scope is 
converted.
   - Or re-collect the issues first — the extractor upserts with 
`OnConflict{UpdateAll: true}`,
     which refreshes `updated_at`, after which a normal incremental run 
reconsiders them.
   
   After that, if anything still looks missing, the PR also adds a diagnostic 
that should tell
   you whether it is scope rather than a bug. It logs, once per board:
   
   > `N collected changelog item(s) are not associated with board <id> and were 
not converted;
   > they belong to issues outside this board's scope.`
   
   Previously those items were silently dropped, so "missing" and "out of 
scope" were
   indistinguishable in the logs. If that warning accounts for your remaining 
delta, the data
   is being excluded by board scope by design and you would need a board 
covering those issues.
   
   Please do report back — if the numbers don't line up after a full refresh I 
would like to
   know, since that would point at something beyond this fix.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to