I appreciate the current stats available, but something I'm trying to determine is how much my current servers overlap.
Would there be any way to record the amount of article availability that overlaps the various servers?
Example:
I have three servers, A, B, and C.
I'd like to see something like:
60% of articles were on all three.
5% were only on A&B
10% were only on B&C
3% were on B only
5% were on C only.
1% were on A only.
16% were not found.
Why?
If I saw the above stats, I'd serious reconsider the need to maintain server A, since it provided so little over having B&C.
More detailed server stats
-
- Newbie
- Posts: 10
- Joined: January 22nd, 2021, 9:05 pm
Re: More detailed server stats
SAB does not know that: SAB tries an article on server X, and if it's there SAB has it, and SAB will not try other servers as no need for that.60% of articles were on all three.
5% were only on A&B
-
- Newbie
- Posts: 10
- Joined: January 22nd, 2021, 9:05 pm
Re: More detailed server stats
Hmm. Okay, that makes sense.
I think I could get the end analysis I'd like with a different bit of bookkeeping, and with that in mind.
I still have servers A,B,C. If I put them all at the same priority, presumably they all get some portion of the first attempt, and then a portion of second and third attempts.
Thus, if we could expose:
20% "Tried A, success"
15% "Tried B, success"
17% "Tried C, success"
5% "Tried A, fail, Tried B, success"
13% "Tried A, fail, Tried C, fail, Tried B, fail"
etc.
I've able to kinda get these numbers by hand, through shuffling their priorities, and then clearing stats, running for a while, and seeing the % completion rates, but this has a couple main shortcomings:
- I'd ideally like them all the same priority for speed
- Clearing stats constantly loses so much history.
I think I could get the end analysis I'd like with a different bit of bookkeeping, and with that in mind.
I still have servers A,B,C. If I put them all at the same priority, presumably they all get some portion of the first attempt, and then a portion of second and third attempts.
Thus, if we could expose:
20% "Tried A, success"
15% "Tried B, success"
17% "Tried C, success"
5% "Tried A, fail, Tried B, success"
13% "Tried A, fail, Tried C, fail, Tried B, fail"
etc.
I've able to kinda get these numbers by hand, through shuffling their priorities, and then clearing stats, running for a while, and seeing the % completion rates, but this has a couple main shortcomings:
- I'd ideally like them all the same priority for speed
- Clearing stats constantly loses so much history.
Re: More detailed server stats
First of all: I'm quite sure nobody is going to implement in SAB what you want.
The good news: with sabnzbd set to +DEBUG, you can analyze anything you want, and try to find all kinds of correlations. Keyword in sabnzbd.log is "missing". Check it out.
PS: SAB 3.2.0 provides info per server. Have you seen that?
The good news: with sabnzbd set to +DEBUG, you can analyze anything you want, and try to find all kinds of correlations. Keyword in sabnzbd.log is "missing". Check it out.
PS: SAB 3.2.0 provides info per server. Have you seen that?
-
- Newbie
- Posts: 10
- Joined: January 22nd, 2021, 9:05 pm
Re: More detailed server stats
Yea, I'm running 3.2. It's what got me looking for more data.
Hmm.
I'll have to ponder the log analysis. Just worry then about the disk space the logs will consume at that level.
Hmm.
I'll have to ponder the log analysis. Just worry then about the disk space the logs will consume at that level.
Re: More detailed server stats
If you're downloading with SAB, I cannot imagine the diskspace for sabnzbd.log will be your bottleneck ...adarkmethodicrash wrote: ↑January 29th, 2021, 9:51 am I'll have to ponder the log analysis. Just worry then about the disk space the logs will consume at that level.
-
- Newbie
- Posts: 10
- Joined: January 22nd, 2021, 9:05 pm
Re: More detailed server stats
I have a machine with all the mechanics, but the *arrs punt the finished product off to various NAS arrays. So not quite the same destination.