feat(eap): Add confidence interval calculations for count #6568
Conversation
|
If I have multiple timeseries in one request. For example a sum, count, avg, p90 How would I know which confidence intervals are for which aggregations? |
snuba/web/rpc/common/common.py
Outdated
| aggregation: AttributeAggregation, | ||
| ) -> list[Expression]: | ||
| """ | ||
| Percentiles require special handling as they require sorting the data by the attribute value |
There was a problem hiding this comment.
and where is this sorting being done?
There was a problem hiding this comment.
still working on it, but it's gonna be done after the results are fetched from clickhouse
There was a problem hiding this comment.
so you're going to sort in-memory of the api server?
There was a problem hiding this comment.
Yeah that was my plan, I could also do it in the query, but the downside is that the query will use more memory. I figured that doing this sorting is gonna slow regardless of where we do it, but doing on the api server will at least reduce overall memory usage in the system. I think we should definitely rethink this in the future though.
Reliability is added for each data point in the time series, and data points directly correspond to a timeseries. I'm not sure if that answers your question. |
davidtsuk
changed the title
❌ 1 Tests Failed:
View the top 1 failed tests by shortest run time
To view more test analytics, go to the Test Analytics Dashboard |
WIP
Closes https://github.com/getsentry/projects/issues/393