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 searched the Issues on Git but couldn't find anything about this. Apologies if I missed it (I really did search - open and closed issues).
When enabling statistics retention for longer than 24h, one can't drill down through the stats to check the server's health in the last (e.g.) 12h or 24h. It's only possible to see stats for the entire period logged (e.g. 90 days). If I want to know whether my server is being performant since X tweak or Y upstream change, I can't (easily) without wiping the current stats log - as only averaged data for the whole logging period (90 days) is shown.
For example, if an AGH instance is being slow for the last 12 hours due to a broken upstream resolver, you can't easily see this in 'Average processing time' on the Dashboard, because that is showing a 90 day average. A mere 12 hours of bad/slow resolving just isn't enough to skew an average of data over such a long period of time. If the user could simply click 'Stats in the last: 12 hours' then instantly the problem would be visible, e.g 'Average processing time in the last 12h: 160ms'.
Proposed Solution
Have a drop-down menu, button or similar to change the stats displayed on the dashboard as (for example) the last 1 hour, 12 hours, 24 hours, week, month etc. That way the user can see quickly and easily whether the server is running at peak health over the last (short period of time) without trying to decipher a 90 day long graph or list of blocked domains.
Alternatives Considered
It's possible to simply retain only 24h of statistics, but this loses a lot of advantages for the user (long-term traffic patterns, ad and tracker behaviour changes over time, etc) to gain a single advantage - knowing the server was performing OK 'today' (rather than 'for the last 90 days - on average).
Additional Information
N/A
The text was updated successfully, but these errors were encountered:
I know however i was making optional suggestion to place time series and data filter control panel to one place(#481 (comment)) for apply changes to all pages, components at once. I remember you said that might be difficult to implement but this issue still duplicate of #314
Problem Description
I searched the Issues on Git but couldn't find anything about this. Apologies if I missed it (I really did search - open and closed issues).
When enabling statistics retention for longer than 24h, one can't drill down through the stats to check the server's health in the last (e.g.) 12h or 24h. It's only possible to see stats for the entire period logged (e.g. 90 days). If I want to know whether my server is being performant since X tweak or Y upstream change, I can't (easily) without wiping the current stats log - as only averaged data for the whole logging period (90 days) is shown.
For example, if an AGH instance is being slow for the last 12 hours due to a broken upstream resolver, you can't easily see this in 'Average processing time' on the Dashboard, because that is showing a 90 day average. A mere 12 hours of bad/slow resolving just isn't enough to skew an average of data over such a long period of time. If the user could simply click 'Stats in the last: 12 hours' then instantly the problem would be visible, e.g 'Average processing time in the last 12h: 160ms'.
Proposed Solution
Have a drop-down menu, button or similar to change the stats displayed on the dashboard as (for example) the last 1 hour, 12 hours, 24 hours, week, month etc. That way the user can see quickly and easily whether the server is running at peak health over the last (short period of time) without trying to decipher a 90 day long graph or list of blocked domains.
Alternatives Considered
It's possible to simply retain only 24h of statistics, but this loses a lot of advantages for the user (long-term traffic patterns, ad and tracker behaviour changes over time, etc) to gain a single advantage - knowing the server was performing OK 'today' (rather than 'for the last 90 days - on average).
Additional Information
N/A
The text was updated successfully, but these errors were encountered: