feat(escalating-v2): Emit escalating metrics #52774
Conversation
github-actions
bot
added
the
Scope: Backend
| org_id=event.project.organization_id, | ||
| project_id=event.project.id, | ||
| metric_name="event_ingested", | ||
| value=1, |
There was a problem hiding this comment.
@ayirr7 this should get aggregated as part of the counter right?
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #52774 +/- ##
==========================================
+ Coverage 79.48% 79.50% +0.02%
==========================================
Files 4935 4940 +5
Lines 207668 208164 +496
Branches 35453 35507 +54
==========================================
+ Hits 165069 165505 +436
- Misses 37570 37621 +51
- Partials 5029 5038 +9
|
| metric_name="event_ingested", | ||
| value=1, | ||
| tags={"group": event.group_id}, | ||
| unit=None, |
There was a problem hiding this comment.
what is unit?
There was a problem hiding this comment.
it's the mri_unit in this template ":/@". I've mostly seen units of time (millisecond, second) here
There was a problem hiding this comment.
@ayirr7 what should this be set to for our needs?
There was a problem hiding this comment.
The unit for the metric you are emitting (whatever you are measuring). If there is no unit then None is fine
src/sentry/tasks/post_process.py
Outdated
| project_id=event.project.id, | ||
| metric_name="event_ingested", | ||
| value=1, | ||
| tags={"group": event.group_id}, |
There was a problem hiding this comment.
Tag keys and values must both be strings to be compatible with the Kafka schema here https://github.com/getsentry/sentry-kafka-schemas/blob/main/schemas/ingest-metrics.v1.schema.json
|
PR reverted: 89889c3 |
This reverts commit 11e749f. Co-authored-by: beezz <1500654+beezz@users.noreply.github.com>
| @@ -16,6 +16,8 @@ | |||
| from sentry.issues.grouptype import GroupCategory | |||
| from sentry.issues.issue_occurrence import IssueOccurrence | |||
| from sentry.killswitches import killswitch_matches_context | |||
| from sentry.sentry_metrics.kafka import KafkaMetricsBackend | |||
There was a problem hiding this comment.
@ayirr7 I think we should hide which backend gets used from the caller.
it should be something like
from sentry.sentry_metrics import get_sentry_metrics_backend
# today this returns a KafkaMetricsBackend, but later it
# should be configurable from somewhere outside the importing file
metrics_backend = get_sentry_metrics_backend() This also makes it easier to only have one connection (pool) per process
There was a problem hiding this comment.
Sentry has the service delegator / LazyServiceWrapper abstraction which is commonly used for this
Update the ESCALATING_ISSUES use case during post_process as we see new events come in.