feat(cells): Add post deploy migration to fix OrganizationMapping.date_created field#115406
Closed
lynnagara wants to merge 8 commits into
Closed
feat(cells): Add post deploy migration to fix OrganizationMapping.date_created field#115406lynnagara wants to merge 8 commits into
lynnagara wants to merge 8 commits into
Conversation
…e_created fields Follow-up to #115325. That PR made `update_organization_mapping_from_instance` carry `Organization.date_added` into `OrganizationMapping.date_created` going forward, but existing mapping rows still hold whatever `timezone.now()` was captured at mapping insert time. This migration repairs them. The migration emits one `CellOutbox` per `Organization` with `category=ORGANIZATION_UPDATE`. The standard cell receiver drains each one and re-runs `Organization.handle_async_replication`, which now plumbs `date_added` → `date_created` through to the mapping. This is a post-deploy migration to be triggered manually. Note: this migration replicates the whole row. `date_updated` wil be bumped on every row. Nothing reads that today (Synapse will, later). One alternative appraoch that I considered was a new temporary, more targeted RPC method that updates only the `date_created` row. I decided to reuse the existing RPC method here - creating all the throwaway methods, receivers, categories is probably not worth it for a one-shot migration, and adds additional risk associated with incorrect implementation.
github-actions
Bot
added
the
Scope: Backend
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
Contributor
|
This PR has a migration; here is the generated SQL for for --
-- Raw Python operation
--
-- THIS OPERATION CANNOT BE WRITTEN AS SQL |
![sentry[bot]](https://avatars.githubusercontent.com/in/12637?s=60&v=4){kind=small}
wedamija
reviewed
May 12, 2026
Comment on lines
+40
to
+42
| for org_id in Organization.objects.values_list("id", flat=True).iterator( | ||
| chunk_size=_BATCH_SIZE | ||
| ): |
Member
There was a problem hiding this comment.
I don't completely remember why, but I feel like .iterator doesn't work for us... I think it tries to use server side cursors maybe, which don't work due to pgbouncer
Instead, you can use something like
for org_id in RangeQuerySetWrapperWithProgressBar(
Organization.objects.values_list(id", flat=True),
result_value_getter=lambda values: values[0],
):
Member
There was a problem hiding this comment.
Alternatively, there aren't that many orgs. You could just bring the whole set of org ids into memory and iterate over them
Member
There was a problem hiding this comment.
Wouldn't using RangeQuerySetWrapperWithProgressBar be a solution we know is going to work?
Member
There was a problem hiding this comment.
It's the more standard solution, I just like to give folks options
markstory
reviewed
May 12, 2026
Member
markstory
left a comment
markstory
left a comment
There was a problem hiding this comment.
Did you consider using the existing outbox backfill tooling for this? By creating this many outboxes all at once, we are more likely to impact latency for customer operations, and potentially contribute to contention in postgres.
Comment on lines
+40
to
+42
| for org_id in Organization.objects.values_list("id", flat=True).iterator( | ||
| chunk_size=_BATCH_SIZE | ||
| ): |
Member
There was a problem hiding this comment.
Wouldn't using RangeQuerySetWrapperWithProgressBar be a solution we know is going to work?
Contributor
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 08b272d. Configure here.
Member
Author
are you saying incrementing the replication version will take care of everything? |
Contributor
|
This PR has a migration; here is the generated SQL for for --
-- Raw Python operation
--
-- THIS OPERATION CANNOT BE WRITTEN AS SQL |
Contributor
Backend Test FailuresFailures on
|
Member
I'm not too familiar with this tooling. If we don't need to have this all run on self hosted maybe it makes more sense? |
lynnagara
added a commit
that referenced
this pull request
May 13, 2026
… date_created (#115423) Triggers backfill_outboxes.py to re-replicate every Organization upserting OrganizationMapping.date_created with the value from Organization.date_added now that #115325 plumbs it through. Alternative to #115406 Roll out: - needs to be flipped to 5 in sentry-options-automator to take effect in production
Member
Author
|
superceded by #115423 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Context
Follow-up to #115325. That PR made
update_organization_mapping_from_instancecarryOrganization.date_addedintoOrganizationMapping.date_createdgoing forward, but existing mapping rows still hold whatevertimezone.now()was captured at mapping insert time. This migration repairs them.Implementation
The migration emits one
CellOutboxperOrganizationwithcategory=ORGANIZATION_UPDATE. The standard cell receiver drains each one and re-runsOrganization.handle_async_replication, which now plumbsdate_added→date_createdthrough to the mapping.Roll out notes
This is a post-deploy migration to be triggered manually.
The migration replicates the whole row. The
date_updatedfield will be bumped on every row. Since nothing reads that today (only Synapse will, later), this should be safe.An increase in memcached set commands are expected during the migration as caches will be updated for each row. Given the number of orgs (relatively small) and the outbox delivery concurrency cap of 5, this volume is trivial for memcached.
Alternative approaches considered
An alternative approach considered was a new temporary, more targeted RPC method that updates only the
date_createdrow and not the rest of the fields. However creating all the throwaway methods, receivers, categories comes with it's own cost and adds additional risk associated with incorrect implementation. It's probably not worth it for a one-shot migration.