New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat(replay): add kafka config for new ingest-feedbacks consumer #66169
Commits on Mar 1, 2024
Commits on Mar 4, 2024
Commits on Mar 5, 2024
Commits on Mar 6, 2024
-
feat(replay): remove organizations:session-replay-event-linking featu…
…re flag (#66257) removes flag declaration and all refs getsentry/team-replay#387
-
fix(metrics-extraction): Fix meta lookup in factory (#66256)
Was using TSResult here which is wrong, it's a dict w/ data or a dict of dicts if it's a group-by. The presence of the data key should be indicative. Fixes SENTRY-2V84
-
fix(metrics-extraction): Allow overriding split decision (#66259)
### Summary If there are any problems with the split decision this will cause it to always re-run the code and save the new decision,
-
-
feat(new-trace): Added logic for red errored rows. (#66194)
<img width="1035" alt="Screenshot 2024-03-03 at 9 36 22 PM" src="https://github.com/getsentry/sentry/assets/60121741/860a6370-b1aa-41dc-b5c8-b62b9778ffa5"> --------- Co-authored-by: Abdullah Khan <abdullahkhan@PG9Y57YDXQ.local>
-
feat(replay): adds current user to replay in header (#66253)
This PR adds the current user to the header of the replay in the issue details page: Before: ![Screenshot 2024-03-04 at 1 15 37 PM](https://github.com/getsentry/sentry/assets/8533851/614047aa-e0da-4520-91d0-5d71f22f8cd9) After: ![Screenshot 2024-03-04 at 1 15 31 PM](https://github.com/getsentry/sentry/assets/8533851/4592b55d-ac83-4448-8af5-06988dfc5aa1)
-
fix(crons): Remove db constraint from historical migrations for monit…
…or environment (#66265) These constraints are causing migrations to fail in getsentry, where we've set up the router to move these tables to a separate database.
-
feat(backup): Enable shimming deleted fields (#66248)
We need to support deleted fields on Django model schemas for a number of self-hosted released (currently 2). This means that even if we delete a field on HEAD, we still need to make the import script aware that the field could exit on imported JSON, and pre-process that JSON in such a way that the field is removed in cases where the JSON was generated by an old version that still contained the field. There is no purpose in testing this code standalone. It passes tests in #65923, which is the first actual use of this shimming mechanism.
-
ref: Move topics to new-style of Kafka topic definition (#66271)
Just moving 3 consumers as a first step, will migrate the rest as a follow up
-
ref(crons): Remove unused code branch in mark_failed (#66284)
failure_issue_threshold will never be falsey due because of this condition here [0] [0]: https://github.com/getsentry/sentry/blob/11e039354141a8487776be19d7535d0da2ee53ce/src/sentry/monitors/logic/mark_failed.py#L44-L45
-
-
Clean up
ArtifactBundle(FlatFile)Index
, step 2 (#66207)This is the second step following #66206: Step 2 of removing outdated tables and columns from the `ArtifactBundleIndex` involves removing them from the migration state *only*. Actual removal from SQL follows in a third step.
-
ref(ddm): Remove normalization logic (#66132)
Remove normalization logic from FE. - closes #65765
-
-
-
fix(ddm): Duplicate chart sample ids (#66301)
* Use spanId for identifying chart samples * Wrap widget with error boundary * Fix query response meta type
-
Remove
FileBlob
locking entirely (#66297)Instead of using a Redis-based lock to protect concurrent `FileBlob` uploads and deletes, the code was already changed previously to handle the cases of concurrent uploads, as well as to reduce the likelihood of concurrent deletes by regularly bumping the `FileBlob` timestamp. The lock has been disabled via an `option` recently, and now it is time to completely remove this code.
-
ref(perf): Introduce
MetricReadout
component (#66254)Extracted from #66229 as a first step. The `MetricReadout` component represents a single metric block in our various Starfish metrics ribbons. This component has some advantages over using plain `Block`s: - consistent loading state - consistent missing data state - built-in support for different kinds of units - built-in simple support for tooltips This makes it easy to use it in all kinds of contexts.
-
Bump Sentry Python SDK to version 2.0.0a2 (#66052)
These are changes necessary to run `sentry` with the Python SDK 2.0.0a2. Three tests are still failing and I commented them out to be able to test this on S4S and Canary. The package `pytest-sentry` is not yet compatible to SDK 2.0, we will release an updated version of it soon. After this has been released we can upgrade `sentry` to SDK 2.0 for good. --------- Co-authored-by: getsantry[bot] <66042841+getsantry[bot]@users.noreply.github.com> Co-authored-by: Francesco Vigliaturo <francesco.vigliaturo@sentry.io> Co-authored-by: ArthurKnaus <arthur.knaus@sentry.io>
-
fix(metrics-extraction): Fix attribute vs dict lookup (#66306)
### Summary Missed that the inner block was also getting `data` as if it was an attribute. Fixes SENTRY-2VPR --------- Co-authored-by: getsantry[bot] <66042841+getsantry[bot]@users.noreply.github.com>
-
Revert "Bump Sentry Python SDK to version 2.0.0a2 (#66052)"
This reverts commit b9b433b. Co-authored-by: antonpirker <202325+antonpirker@users.noreply.github.com>
-
fix(performance): Fix a typo in the mri of measurements.inp (#66273)
`measurements.inp` should point to `d:spans/webvital.inp@millisecond`
-
-
Revert "ref(sdk): Bump js sdk version to 7.105.0 (#66224)"
This reverts commit e968207. Co-authored-by: edwardgou-sentry <83961295+edwardgou-sentry@users.noreply.github.com>
-
fix(profiling): Add missing transaction condition for profile digest (#…
…66309) This was querying all profiles accidentally without the transaction condition.
-
feat(proguard): Add Symbolicator.process_jvm method (#66064)
This is part of the effort to port proguard processing to Symbolicator (getsentry/team-processing#116). In particular, it hooks into the new `symbolicate-jvm` endpoint introduced in getsentry/symbolicator#1393.
-
fix(hybridcloud) Add visiblity to region configuration (#66295)
When we add regions we need a way to keep the region out of UI interactions and the public API responses until it has been validated. The `find_all_multitenant_region_names` method is used in sentry and getsentry to get the list of multi-tenant regions that are available for new account creation. Once we have a new region functionally working it can be made visible. Even after a region is visible account creation would be gated by feature flags.
-
Just a quick padding fix on the resource page, nothing fancy. Before ![image](https://github.com/getsentry/sentry/assets/44422760/ff0097a0-5309-43e1-bea3-744daf827e28) After ![image](https://github.com/getsentry/sentry/assets/44422760/4e0adbfc-87f6-48cd-bb6a-9acae370a3dc)
-
feat(sdk): bumps js sdk to 7.103.0 and updates browsertracingintegrat…
…ion (#66317) bumps js sdk to 7.103.0 and updates browsertracingintegration
-
feat(indexed-spans): Support profile.id on indexed spans (#66178)
This column is called `profile.id` on other datasets. Not sure why it's called `profile_id` here. Especially when all other id columns in this dataset uses the dot notation. Keeping `profile_id` around for compatibility. --------- Co-authored-by: William Mak <william@wmak.io>
-
fix(perf): Fix HTTP widget documentation link (#66011)
Fixes #65071 Directs users with no HTTP span to the "Set Up Performance" page rather than the "Integrations" page which doesn't even exist for a lot of platforms.
-
feat(processing) Move redis operations in reprocessing to servicewrap…
…per (#66168) We need to shift the reprocessing workload to a new redis cluster without downtime. In order to do that I would like to use our Service wrappers and ServiceDelegator to dual write and then shift reads via runtime options. Before I can move load with ServiceDelegator the existing logic needs to be moved into a service wrapper.
-
ref(metrics): Change aggregation options to be a dictionary (#65666)
Change the get_aggregation_options to return a dictionary of options, and associate the options with a time window. This is precursor work to being able to check the aggregation options when querying, to ensure queries are not asking using options that are not available.
-
ref(crons): Remove legacy fingerprint generation (#66250)
While implementing monitor incidents there was a period of time where we were "cutting over" to monitors creating incident entities instead of just producing issue occurrences where the grouphash was computed from an interim 'last_state_change'. All monitors are now producing incidents when they fail, so we no longer need the fallback grouphash computation The `last_state_change` was only required for this interim period and will also be removed in a future commit.
-
feat(issue-priority): Rename issue stream tab inbox -> priority and r…
…emove unresolved tab (#66269)
-
ref: Move more code to the new way of defining kafka topics and overr…
…ides (#66283) This moves the `dlq_topic` consumer definitions and anything that uses the `get_topic_definition` helper function over to the new style kafka topic config.
-
feat(performance): Updates useIndexedSpans to accept an optional fiel…
…ds array (#66275) Updates useIndexedSpans to accept an optional fields array to allow selecting specific fields. Refactors existing call to use new function signature.
-
-
feat(issue-priority): Use 'Med' label in priority pills (#66272)
Also adjusts the stream column width since we don't need as much space any longer
-
styles(metrics): Use disabled button for no profile (#66341)
The no value message is quite long so use a disabled button for it to appear more compact.
-
ref(feedback): update links and system default for onboarding (#66330)
Changed snippet example to `colorScheme: "system"`: <img width="457" alt="SCR-20240305-jdjj" src="https://github.com/getsentry/sentry/assets/56095982/fe99f6aa-8b3e-4ed6-b117-15cacf9273dd"> Updated links and hyperlink to the specific "bring your own button" section in the docs: <img width="465" alt="SCR-20240305-jcve" src="https://github.com/getsentry/sentry/assets/56095982/2a9984a4-4947-437c-bcc5-b93d991a70b9">
-
chore(commit-context): Remove unused feature flag (#66345)
With #65346 and getsentry/getsentry#12983 merged, self-hosted and all organizations on SaaS now have access to the commit-context feature for suspect commits. The feature flag can be safely removed.
-
ref(crons): Remove usage of last_state_change (#66328)
This code is no longer needed now that we have stopped generating legacy event hashes in GH-66250
-
fix(metrics): Update indexer telemetry to handle no messages (#66350)
### Overview Update indexer telemetry to handle no messages in case everything is killswitched
-
fix(ui): Remove custom prism text selection styles (#66240)
The existing styles were too similar to the highlighted line color. Using browser default coloring works well enough.
-
ref(feedback): add rust to platforms (#66356)
<img width="459" alt="SCR-20240305-lrbx" src="https://github.com/getsentry/sentry/assets/56095982/90706ed5-cf5d-4996-a9e1-79b94f38e7d2">
-
Revert "ref: Move more code to the new way of defining kafka topics a…
-
Revert "ref: Move query subscriptions over to new-style kafka config (#…
-
ref(crons): Drop last_state_change column (#66353)
This is no longer needed after usage removal in GH-66328
-
bump more_slow_alerts to 300 alerts (#66367)
[per discussions](https://sentry.slack.com/archives/C06J6HKFKLG/p1708464708399359) bumping the 'more slow alerts' group to 300 alerts. This is a temporary fix and will need to be addressed properly (potentially via [inc-666](https://getsentry.atlassian.net/browse/INC-666), but there may be other solutions here as well)
-
Activated alert rule apis feature flag (#66361)
Includes feature flag and monitor type validator in the alert rule api serializer Serializer converts int value into enum
-
ref(daily summary): Add logger when sending summary (#66369)
Add a log line when we send the notification as I'm doing internal testing.
-
ref(crons): Clean up mark_failed logic (#66372)
Just some minor re-organiztion to make it a bit more readible
-
docs: Wrap stories with OrganizationContainer (#66323)
<img width="687" alt="SCR-20240305-ihne" src="https://github.com/getsentry/sentry/assets/187460/e15745d8-ea24-4ce6-a2ec-858a6c50d1b8">
-
logging(group-owner): adds logging for create group owners (#66371)
Used for debugging race conditions of when we create the row vs when we check for it
-
fix(crons): Fix backfill script to handle the case where a monitor wi…
…th a slug already exists in the target org (#66340) There are 4 monitors that exist twice in the same project, but in different orgs. The ones that have a different org than the project's org are totally broken and can't receive checkins. We'll just delete them, they're inaccessible at the moment.
-
ref(feedback): remove duplicate native platform (#66384)
accidentally had two `native` platforms in this list
-
feat(issue-details): Improved UI for viewing object/array values (#66153
) Fixes #58573, #48480 Previously, objects/arrays were only collapsible after they reached maxDefaultDepth. The major change here is that now each level is collapsible - Each level is now always collapsible. Before, objects/arrays were only collapsible after a certain depth - Arrays/objects begin collapsed when there are more than 5 items, in addition to when they are at a certain depth - Changed the appearance and location of the toggle buttons to better match commonly-used json viewers - Added `n items` preview text for collapsed objects/arrays (which is clickable!)
-
ref: Move code to new-style kafka topics - take 2 (#66381)
Brings back #66283 with fixes to the indexer strategy. It crashed on the prior deploy as we were not producing to the correctly resolved topic.
-
chore(crons): Return a better message from deprecated crons ingest ap…
…is (#66388) <!-- Describe your PR here. -->
-
ref(ddm): Remove multiple metrics meta queries (#66305)
Remove request duplication logic for metrics meta queries to utilize parallelization on the BE and unblock other requests in the browser.
-
-
Bump Sentry Python SDK to version 2.0.0a2 (#66397)
New PR for another load test to check performance of the new Python SDK 2.0 --------- Co-authored-by: getsantry[bot] <66042841+getsantry[bot]@users.noreply.github.com>
-
Revert "Bump Sentry Python SDK to version 2.0.0a2 (#66397)"
This reverts commit 51ae02e. Co-authored-by: antonpirker <202325+antonpirker@users.noreply.github.com>
-
feat(ddm): Filter empty series from summary table (#66400)
Workaround as the endpoint returns empty series for groups that are excluded by a filter.
-
-
feat(ddm): Hiding queries (#66398)
Allow toggling queries from being displayed by clicking their query symbol. - closes #66201
-
-
ref(deps): Bump sentry-relay to v0.8.48 (#66401)
Along the way of bumping dependencies, some tests are fixed to cover for the changes in Relay.
-
chore(hybridcloud) Remove unused parameter from sent_incident_alert_n…
…otification (#66113) Complete work started in #65736, #65791, and #65883 to remove a parameter from the send_incident_alert_notification RPC method. Fixes HC-1123 Fixes SENTRY-2QAG
-
fix(hybridcloud) Make bitbucket uninstall silo safe (#66354)
Follow up on a TODO that has triggered an error. Add tests so that future maintenance on the bitbucket integration is a bit easier. Fixes SENTRY-2TNS
-
fix(hybridcloud) Add logging for bad RPC requests (#66357)
Add logging when we have a bad request from an RPC method. We don't often get sentry errors from these failures, and not being able to see the response of these operations makes debugging harder.
-
feat(replay): Add feature flag for mobile replay player (#66276)
This adds a feature flag to allow support of mobile replay player. Currently for internal testing only
-
feat(spans): Tag perf problems created from standalone spans (#66325)
As we set up the application logic to run performance issue detection on standalone spans, we want to collect metrics to check if transaction based and span based perf issue detection are equivalent. Tagging metrics with `is_standalone_spans` to track this. This also updates fingerprint so the standalone span occurrences don't get grouped with existing perf issues during testing.
-
Remove rollout options for Rust Enhancers (#65966)
We have fully enabled these, so lets remove all the options usage related to it. This is pretty much a followup from #65533.
-
feat(profiling): pass internal (profile-scoped) dsn for function metr…
…ics ingestion to vroom (#66300)
-
fix(mql): Handle strings and formula with scalars bug (#66324)
This PR fixes two bugs. Firstly, if a user passed an MQL string into the layer (which is completely valid) the layer would throw an exception. Secondly, if a formula had a component that was itself a formula of scalars, e.g. sum(x) + (7 * 24), an exception would be thrown.
-
feat(hybridcloud) Add region tag to errors/transactions (#66370)
Add sentry_region to the tags of errors/transactions we collect. This helps with diagnosing which region an error is occurring in. Previously we were planning on using `environment` for this but it would have required a non-trivial amount of monitoring changes. The `sentry_region` name mirrors the tag used in other monitoring tools we use.
-
feat(perf): Add HTTP module domain overview page scaffold (#66359)
Just some very very basic content to get some content and routing up.
-
-
ref(onboarding): refactor useCurrentProjectState to be generalized (#…
…66118) - DDM, Replays, and Feedback onboarding were all using some variation of this hook: `useCurrentProjectState`. They're all pretty similar, so I refactored them into one general reusable hook - I modified the respective `sidebar.tsx` files to use this new hook - I also added a hook test
-
Revert "ref(onboarding): refactor useCurrentProjectState to be genera…
-
feat(code-mappings): LA automatic code mapping support for PHP projec…
…ts (#66151) (This feature is in LA) Whitelisted php projects for automatic code mapping support and added unit tests --------- Co-authored-by: Bartek Ogryczak <bartek.ogryczak@sentry.io> Co-authored-by: Armen Zambrano G. <44410+armenzg@users.noreply.github.com>
-
-
ref(onboarding): refactor useCurrentProjectState to be generalized (#…
…66425) - had to revert the original PR (#66118) due to a failing test - fixed in this PR! this is the only change from the previous PR - DDM, Replays, and Feedback onboarding were all using some variation of this hook: `useCurrentProjectState`. They're all pretty similar, so I refactored them into one general reusable hook - I modified the respective `sidebar.tsx` files to use this new hook - I also added a hook test
-
fix(metrics): Adjust bounding boxes based on operation (#66339)
The bounding boxes right now always bound the metric on the min/max. But depending on the operation, we should look at different values to bound determine if it lies in the bounding box. For example, if graphing the min, we should ensure the min is with in the bounding box.
-
chore(superuser): Add loggers to _needs_validation (#66364)
There's a bug where we're no longer requiring superuser access category and reason validation, which is caused by https://github.com/getsentry/sentry/pull/66043/files. I want to validate what those booleans are in order to find out if this is the cause of the bug.
-
feat(discover): allows filtering metric queries on span.op (#66423)
Adds `span.op` to `DEFAULT_METRIC_TAGS` so that it is available in discover metrics queries.
-
ref(replay): see full replay in rage click issue goes to breadcrumbs …
…tab (#66374) From a rage click issue, "see full replay" takes you to the breadcrumbs tab with the rage & dead click filter selected. Relates to getsentry/team-replay#394. Temp solution until we can load in rage click issues to the errors tab https://github.com/getsentry/sentry/assets/56095982/cda9511d-8458-4fc4-a506-0b36ed217d51
-
ref(perf): Remove usage of generic-ish
SpanMetricsRibbon
(#66415)Another refactor peeled away from #66229. The `SpanMetricsRibbon` component is used in two places: 1. The top right of the Query Summary page 2. The completely unused generic span summary page The ribbon is a little silly because it supports just these two use cases! This is much better done using the new `MetricReadout` by using it directly in the parent. No visual changes.
-
ref: Remove old style topic definition from ingest consumer code (#66352
-
feat(feedback/issues): add sidebar to issue details UF empty state (#…
…66351) - add a sidebar popup when "set up now" is clicked - add crash report modal onboarding config for javascript with widget callout <img width="1072" alt="SCR-20240305-ocgr" src="https://github.com/getsentry/sentry/assets/56095982/a9b92ffb-7ea0-4bd7-83b0-299e3ae2a2fe">
-
fix(feedback): correct usage of option name (#66430)
- I messed up the name of the project option in the endpoint, it should be plural, not singular - I was using the wrong project option flag in the check 🤦♂️ Thank you to @jas-kas for catching this bug. 🙏🏼 this is what EA is for! I do want to refactor project options to make it easier to test // less prone to bugs when adding additional ones. will add a TODO item.
-
ref(replays): remove rage click creation celery task (#66360)
We've rolled out rage click_creation_with_replay_event, so we can remove this celery task and the usages of it. huzzah! https://sentry.sentry.io/dashboard/77863/?project=1&statsPeriod=24h (can see bottom metric has dropped to zero, while one to left has gone up). this represents that now the new function is being called 100% of the time so we can remove this code.
-
fix(queries): Add more errors to queryhandler (#66432)
- Add urllib3 errors to the query handler - Mostly downstream hiccups in either case - Fixes SENTRY-2W3A - Fixes SENTRY-2TE9 - Fixes SENTRY-2VVR - Fixes SENTRY-2JBW
-
fix(profiling): Add handle_query_errors to profiling (#66433)
- This was missing a handle_query_errors so exceptions were bubbling up too high - Fixes SENTRY-2KVM
-
feat(crons): Add new broken detection model (#66375)
Spun off from: #65963
-
test(ui): Switch children to react nodes (#66390)
- adds a missing children prop, swaps ReactElement to ReactNode - removes a render function used as a child
-
-
feat(crons): Show project information in details area for crons (#65944)
As we transition to a state where monitor slugs are unique to the project and not the org, we need to have project slug information in the crons listing page to differentiate duplicate cron slugs within an org. <img width="1246" alt="image" src="https://github.com/getsentry/sentry/assets/9372512/f31e4f75-cdc0-41d3-820e-a2928562fe73"> If the project + schedule text gets too long, then it wraps to the next line. We have a good amount of space in the details area so opted for wrapping vs. ellipsizing
-
-
ref(crons): Drop last_state_change from db (#66429)
No longer needed after GH-66353
-
feat(webvitals): Updates web vital module query hooks to actually que…
…ry for INP measurements (#66421) Updates all metrics based queries in the Web Vitals module to properly query for INP measurements and aggregates. We previously subbed in FID metrics for INP for testing/development purposes. This PR removes that since INP is now properly queryable in Discover.
-
feat(webvitals): update calculatePerformanceScoreFromStoredTableDataR…
…ow to properly account for weights from different score profiles (#66434) Update calculatePerformanceScoreFromStoredTableDataRow to properly account for weights from different score profiles
-
feat(metrics): Support paginating samples via random seed (#66440)
To give the illusion of a consistent paging, we take the hash of the limit and offset as the seed to `groupArraySample`.
-
feat(webvitals): Greys out FID in the performance score breakdown cha…
…rt and updates tooltip to label as deprecated (#66438) Greys out FID in the performance score breakdown chart and updates tooltip to label as deprecated
-
feat(webvitals): updates webvitals landing page slideout panel to sup…
…port displaying inp (#66439) updates webvitals landing page slideout panel to support displaying inp
-
ref(sdk): Bump js sdk to 7.105 again (#66436)
### Summary We're bumping again since we've determined that pageloads are now lower due to inheriting the sampling decision from the sentry project, along with the release, public key and transaction name potentially changing
-
fix(event_manager): catch kafka exceptions in perf issue production (#…
…66447) catch KafkaExceptions here, as we shouldn't crash the task. Can follow up with diagnosing the KAFKA_TOO_LARGE payloads and making an upstream fix. Should also determine if we should catch this logic in the occurrence production code instead, so other upstream places don't crash because of this. FIxes https://sentry.sentry.io/issues/5040716121/
-
ref: split incident and alert_rule models into separate files (#66386)
currently the incidents/models.py file is nearly 1k lines long and the file name `model` does not do much to tell us what it's contents are PR to split up the file names and directory edit: oh lawd. i swear most of these are modifying imports now that the directory structure has been modified Resolves getsentry/getsentry#13170 Requires getsentry/getsentry#13170 --------- Co-authored-by: getsantry[bot] <66042841+getsantry[bot]@users.noreply.github.com>
-
fix(ondemand): Wrong column usage here was breaking (#66437)
- If fields had `-` in them the query would fail since we weren't resolving the column to a tag correctly - Fixes: SENTRY-2W5S
-
fix(event-stats): Remove error from top events (#66455)
- Fixes SENTRY-2VV0 - Fixes a bug where we were saying errors supports top events even though it doesn't
-
feat(api): Allowlist for API Pagination Check (#66463)
* Just the allowlist part of the API Pagination Check * Need to break down the larger PR because adding `requires` in the PR description doesn't seem to be working
-
ref(crons): Move grouphash generation to a field default (#66456)
Let's centralize this here, no need for callers to need to think about this detail --------- Co-authored-by: getsantry[bot] <66042841+getsantry[bot]@users.noreply.github.com>
-
ref(crons): Clarify why incident creation skip happens (#66462)
Guard statments tend to be clearer
-
chore(inbound-filters): Remove feature flag for default inbound filte…
…rs (#65989) this pr removes the feature flag for default inbound filters for new javascript projects now that it is GA'ed
-
ref: Remove references to deprecated settings (#66355)
settings.KAFKA_OUTCOMES and settings.KAFKA_OUTCOMES_BILLING are deprecated and will be removed soon
-