Incident impact availability reporting

1.How do I fill in the missing weeks where there is not any data, and just fill that week as 100%
2. When subtracting from 100.00 the precision drops to only 2 places to the right of the decimal, when I want precision of 5 places.

select sum(durationSeconds/6.048e+5*100) as ‘impact to portal percentage’
from NrAiIncident
where targetName = ‘this target’
and conditionName = ‘2 locations failed.’
and priority = ‘critical’
since 12 months ago
facet WeekOf(timestamp) as ‘Week’

Hey @Ronald.Johnson1

You can’t retroactively substitute empty weeks of event data. So for Q1 that’s not going to be possible.

For Q2, you can absolutely change the level of precision though:

See on the right hand side within the Data Format section

1 Like

That’s great ! is there anyway to join the data from the incident table with the synthetics table?
Then I can compute the difference between real location availability, and multiple location availability.

With great difficulty, perhaps. That’s not an easy query to put together, but I’d recommend checking out filter queries, which I think may help.

Filter queries allow you to bring WHERE clauses into the select statement - which will help you to target separate events/attributes.

Here’s a simple example looking at percentage of Transaction Errors as compared to Synthetics success rate.

SELECT filter(percentage(count(*), where error is true), WHERE eventType() = 'Transaction'), filter(percentage(count(*), where result = 'SUCCESS'), WHERE eventType() = 'SyntheticCheck') FROM Transaction, SyntheticCheck SINCE 1 WEEK AGO 

The result of that query is:

Well I will take a look at filters. But the goal here is to find when 2 locations for a web portal are failing at the same time. That is my alert condition to catch that. The alert condition rules out a single location failure. Having 2 locations fail on geographically each side of the datacenter along the internet backbone where the app is running gives us the app uptime, and the outage is seconds from they incident duration.
in the case of multiple incidents within a given faceted time period you have to change the equation a bit like so.

100 - (sum(durationSeconds)/6.048e+5*100) as ‘EACH WEEK AVAILABILITY WAS AFFECTED.’
with a facet of each week.
from NrAiIncident
and conditionName = ‘2 locations failed.’
and priority = ‘critical’ since 12 months ago facet WeekOf(timestamp) as ‘Week’

Keeping single location failure percentages for one synthetic is pretty easy.
Using scripted synthetics gives us a better picture than just counting transactions.

Thanks for your help Ryan :smiley: