Data collector jobs (relative start, breakdowns)

[oana5]

Hi everyone,

I wanted to check with you what do you think it would be better to set as relative start on a daily data collector job?

* I would like to have some historical info, but I don't want the system to capture too many records.

* Also, just to confirm, if we don't tick the checkbox for collecting breakdown matrix inside the indicator, we will not be able to filter the data, right?

Can you please share your thoughts?

Best regards,

Oana

[Erik Hoffmann]

Hi Oana,

Typically the relative start for a daily collection job is 1, and it will collect for one day. If you set if higher it will collect for more periods back. The latter is typically just if you want to overwrite already collected scores. There are some use cases that require this.

If you do not tick 'collecting breakdown matrix' the data collector job will not collect combinations of breakdowns. So if you have 'priority' and 'state' defined as breakdowns it will always collect scores for 'priority' and 'state' (so you can filter on priority = x, or filter on state = y), and if 'collecting breakdown matrix' is set to true it will /also' collect combination of 'priority' and 'state' (so you can filter on priority =   x and state = y).

Best regards,

Erik

[oana5]

Thank you for your reply, Erik.

Yes, it makes perfect sense.

Just to check if my understanding is good - Let's say I want to store monthly logged and closed tickets (incidents, requests). For this, I might have 2 options:

* I collect daily scores for newly logged requests, (daily data collector , relative start 1 day, relative end 1 day) and then create an indicator to represent them by month sum +

* I could just use a data collector to collect data from the past (daily data collector, relative start 1 months ago, relative end 1 day)?

Please correct me if I am wrong.

[Erik Hoffmann · Accepted solution]

Hi Oana,

I don't think we see a lot of monthly values collected directly. Usually daily values are collected and then using the aggregation by sum is more powerful, because it allows also using other aggregations like 7 day sum.

I think in the second option you will need to have an indicator with conditions querying this month, and then run it daily. But not with the relative start to 1 month ago, just 1 day ago. As such every day the sum will increase by the new incidents of the past day -- a running sum of the month. The relative start is used to collect "in the past" where relative start sets the time range of how far back. If you want to collect 'new incidents' of the last 30 days, you would set relative start to 30.

Best regards,

Erik