[ARCHIVED] Peculiar drop in requests

[jeff_longland]

Relatively new to Canvas Data, so pardon my ignorance if this has come up previously (though I didn't find anything in my search). We created a report on requests by hour and day of week. When looking at it, we see a drop in requests every day of the week 4-5pm (Pacific Time).

I haven't dug into this in any detail, but figured I'd ask here. I know the requests table is generated from log events and isn't guaranteed to be accurate, just wondering if there are known issues with particular time periods missing data?

Thanks,

Jeff

[sam_mcknight]

What is the date range you are covering?

I'm sure you've already done this but I just wanted to confirm that you adjusted your time field to PDT and didn't simply make the column header say PDT.

For future question visitors, all times in Canvas Data are GMT.

[jeff_longland]

Re: timezones - The above screenshot is UTC adjusted to PDT.

Looking more closely at the data, it appears that we don't have any requests after 23:17 UTC every day.

date_trunc	count
2017-11-15 23:01:00	4188
2017-11-15 23:02:00	4128
2017-11-15 23:03:00	3220
2017-11-15 23:04:00	3413
2017-11-15 23:05:00	3709
2017-11-15 23:06:00	2970
2017-11-15 23:07:00	3163
2017-11-15 23:08:00	3761
2017-11-15 23:09:00	3234
2017-11-15 23:10:00	3070
2017-11-15 23:11:00	2795
2017-11-15 23:12:00	3173
2017-11-15 23:13:00	2525
2017-11-15 23:14:00	2821
2017-11-15 23:15:00	2505
2017-11-15 23:16:00	2840
2017-11-15 23:17:00	164

I suspect there's something amiss with Canvas Data. I believe we're the first client using Canvas Data in the AWS Canada region - so perhaps it's related. I imagine other clients would have discovered this otherwise.

[a1222252]

Hi Jeff,

We download Canvas data using the canvasDataCli tool. Back in early May we had the same issue. The maximum request timestamp for each day dropped progressively from 23:59:59 to around 23:17:02. I raised a service request and it was fixed by early June.

Our requests data has started to cut off early again, but it's only 10 minutes this time and I haven't got round to raising another service request.

From memory this was to do with how the data was being extracted from the web server access log. I can find the Instructure case number if it helps.

Regards,

Stuart.

[a1222252]

My Instructure case number for this issue was 02088394, assigned on 5th June. It was resolved on 9th June.

[jeff_longland]

Thanks Stuart, much appreciated. Will follow-up.