perf(issues): improve adjacent_events query #79365
Merged
Conversation
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
github-actions
bot
added
the
Scope: Backend
armenzg
approved these changes
Oct 18, 2024
| Condition( | ||
| Column(DATASETS[dataset][Columns.TIMESTAMP.value.alias]), | ||
| Op.LT, | ||
| event.datetime + timedelta(days=100), |
Member
Author
There was a problem hiding this comment.
sentry/src/sentry/eventstore/snuba/backend.py
Line 577 in 55415ea
harshithadurai
pushed a commit
that referenced
this pull request
Oct 19, 2024
Right now the a `adjacent_events` query does a really big scan in order to find the adjacent events given an event id. It does this because the old query does not specify an order by which can take advantage of the primary key `ORDER BY (project_id, toStartOfDay(timestamp), primary_hash, cityHash64(event_id))` This PR adds a new function `get_adjacent_event_ids_snql` which uses SnQL and adds a order by clause. We observe in the snuba admin tool this new query takes only 2 seconds and scans hundreds of MBs of data as opposed to the old query which scans 10+ GB on large issues. `get_adjacent_event_ids_snql` is not general purpose as its only used in one place. see ticket below for more info: Fixes getsentry/team-issues#42
jan-auer
added a commit
that referenced
this pull request
Oct 21, 2024
* master: (288 commits) feat(metrics): Register MRI for spans/count_per_root_project (#78992) feat(dynamic-sampling): Settings for sample rate (#79341) Revert "feat(sentry-sdk): Enable HTTP2 transport" (#79391) fix(feedback): keep oldest date_added for duplicate user reports (#79387) chore(issue-stream): Remove tooltip for Unhandled (#79385) chore(autofix): Show banner if gen AI consent is given, even if no feature flag (#79362) chore(autofix+copilot) Allow autofix without FF if gen AI consent given (#79361) Fixes VULN-50 by enforcing option (#79384) perf(issues): improve adjacent_events query (#79365) feat(issues): Add anchor links back to issue sections (#79333) fix(issue-views): Make tab bar take up entire row (#79383) chore(issues): Add additional metrics for ownership matching (#79302) feat(insights): create screen rendering module (#79192) fix(issues): Avoid streamline issue layout rerenders (#79327) ref(performance): Add missing types to performance widgets (#79301) chore(issue-views): Add translation wrapper to aria label (#79320) chore(issue-stream): Reduce font size of title and message (#79378) feat(insights): update headers and breadcrumbs on frontend domain view (#78945) feat(insights): add view trends button to ai overview (#78611) ref(rr6): Remove unused param (#79379) ...
cmanallen
pushed a commit
that referenced
this pull request
Oct 23, 2024
Right now the a `adjacent_events` query does a really big scan in order to find the adjacent events given an event id. It does this because the old query does not specify an order by which can take advantage of the primary key `ORDER BY (project_id, toStartOfDay(timestamp), primary_hash, cityHash64(event_id))` This PR adds a new function `get_adjacent_event_ids_snql` which uses SnQL and adds a order by clause. We observe in the snuba admin tool this new query takes only 2 seconds and scans hundreds of MBs of data as opposed to the old query which scans 10+ GB on large issues. `get_adjacent_event_ids_snql` is not general purpose as its only used in one place. see ticket below for more info: Fixes getsentry/team-issues#42
JoshFerge
added a commit
that referenced
this pull request
Oct 29, 2024
fixes the adjacent queries introduced in #79365 -- needed to adjust the conditions a bit. adds a new test to ensure correct behavior.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Right now the a
adjacent_eventsquery does a really big scan in order to find the adjacent events given an event id. It does this because the old query does not specify an order by which can take advantage of the primary keyORDER BY (project_id, toStartOfDay(timestamp), primary_hash, cityHash64(event_id))This PR adds a new function
get_adjacent_event_ids_snqlwhich uses SnQL and adds a order by clause. We observe in the snuba admin tool this new query takes only 2 seconds and scans hundreds of MBs of data as opposed to the old query which scans 10+ GB on large issues.get_adjacent_event_ids_snqlis not general purpose as its only used in one place.see ticket below for more info:
Fixes https://github.com/getsentry/team-issues/issues/42