You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been doing some experimentation logging calls to CF Analytics Engine, and observe that the same queries can sometimes take < 1s but also > 10s or even worse (I've observed ~25 seconds sometimes).
Here's an example of the same query SQL but performed on different columns (e.g. blob1, double2, etc) can yield wildly different query times:
query path true OK took 794ms
query country true OK took 903ms
done referrer true OK took 3143ms
done deviceModel true OK took 3325ms
done browserName true OK took 12746ms # <-- 12.7 seconds
Example of one of these queries:
SELECT blob7,
SUM(_sample_interval) AS count
FROM metricsdataset
WHEREtimestamp> Now() - INTERVAL '7' day
AND double1 =1AND blob8 ='counterscale-dev'
GROUP BY blob7
ORDER BY count DESCLIMIT10
I'd personally love to instrument this code with @getsentry and emit metrics/spans etc for individual queries, but the CF <> Sentry integration only captures errors via tail workers.
The text was updated successfully, but these errors were encountered:
I've been doing some experimentation logging calls to CF Analytics Engine, and observe that the same queries can sometimes take < 1s but also > 10s or even worse (I've observed ~25 seconds sometimes).
Here's an example of the same query SQL but performed on different columns
(e.g. blob1, double2, etc)
can yield wildly different query times:Example of one of these queries:
I'd personally love to instrument this code with @getsentry and emit metrics/spans etc for individual queries, but the CF <> Sentry integration only captures errors via tail workers.
The text was updated successfully, but these errors were encountered: