[Nagios-devel] Antwort: Re: Request for comment: Overhaul of
Posted: Wed Apr 02, 2008 5:32 am
Dies ist eine mehrteilige Nachricht im MIME-Format.
--=_alternative 004A60DDC125741F_=
Content-Type: text/plain; charset="utf-8"
content-transfer-encoding: quoted-printable
[email protected] schrieb am 02.04.2008 14:55:15:
> While I agree this particular info is missing, the active check latency
> tells you how much you're lagging behind. As long as it's near 0 you're
> OK, but when it starts to show big latencies (i.e. >30 seconds) even
> keeps growing then you definitely have a problem.
Yes I know. But the main point was, that the presented percentages
are more or less useless. The check latency is by far the only real
relevant information on that page. A fact, that makes the page seem
to be broken by design to me.
> Also note that service latency is not always a "server performance
> problem". In Nagios 2.0, host checks are run serially so the time in
> takes to run host checks can cause latency to grow up pretty fast,
> especially when you start seeing hosts down.
>
> There's other reason for high latencies: perfdata commands, ocsd
> commands, etc. Ideally you can avoid this by having Nagios write a file
> or FIFO instead... There's many example out there.
I do know, but that has actually nothing to do with my initial proposal
This thread was not meant for debugging my performance issues - especially
since I already solved them last week. ^^ It is more or less just a=20
reflection
of the fact, that the Performance Info was just very useless for actually
finding out infos about the performance.
S
--=20
Sascha Runschke
Netzwerk- und Systemmanagement
Telefon : +49 (201) 102-1879 Mobil : +49 (173) 5419665 Fax : +49 (201)=20
102-1102105
GFKL Financial Services AG
Vorstand: Dr. Peter J=C3=A4nsch (Vors.), J=C3=BCrgen Baltes, Dr. Till Ergen=
zinger, Dr. Tom Haverkamp
Vorsitzender des Aufsichtsrats: Dr. Georg F. Thoma
Sitz: Limbecker Platz 1, 45127 Essen, Amtsgericht Essen, HRB 13522
--=_alternative 004A60DDC125741F_=
Content-Type: text/html; charset="utf-8"
content-transfer-encoding: quoted-printable
[email protected] schrieb
am 02.04.2008 14:55:15:
> While I agree this particular info is missing, the active check latenc=
y
> tells you how much you're lagging behind. As long as it's near 0 you'r=
e
> OK, but when it starts to show big latencies (i.e. >30 seconds)
even
> keeps growing then you definitely have a problem.
Yes I know. But the main point was, that the present=
ed
percentages
are more or less useless. The check latency is by
far the only real
relevant information on that page. A fact, that makes
the page seem
to be broken by design to me.
> Also note that service latency is not always a "server performanc=
e
> problem". In Nagios 2.0, host checks are run serially so the
time in
> takes to run host checks can cause latency to grow up pretty fast,
> especially when you start seeing hosts down.
>
> There's other reason for high latencies: perfdata commands, ocsd
> commands, etc. Ideally you can avoid this by having Nagios write a
file
> or FIFO instead... There's many example out there.
I do know, but that has actually nothing to do with
my initial proposal
This thread was not meant for debugging my performan=
ce
issues - especially
since I already solved them last week. ^^ It is more
or less just a reflection
of the fact, that the Performance Info was just very
useless for actually
finding out infos about the performance.
S
--
Sascha Runschk
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]
--=_alternative 004A60DDC125741F_=
Content-Type: text/plain; charset="utf-8"
content-transfer-encoding: quoted-printable
[email protected] schrieb am 02.04.2008 14:55:15:
> While I agree this particular info is missing, the active check latency
> tells you how much you're lagging behind. As long as it's near 0 you're
> OK, but when it starts to show big latencies (i.e. >30 seconds) even
> keeps growing then you definitely have a problem.
Yes I know. But the main point was, that the presented percentages
are more or less useless. The check latency is by far the only real
relevant information on that page. A fact, that makes the page seem
to be broken by design to me.
> Also note that service latency is not always a "server performance
> problem". In Nagios 2.0, host checks are run serially so the time in
> takes to run host checks can cause latency to grow up pretty fast,
> especially when you start seeing hosts down.
>
> There's other reason for high latencies: perfdata commands, ocsd
> commands, etc. Ideally you can avoid this by having Nagios write a file
> or FIFO instead... There's many example out there.
I do know, but that has actually nothing to do with my initial proposal
This thread was not meant for debugging my performance issues - especially
since I already solved them last week. ^^ It is more or less just a=20
reflection
of the fact, that the Performance Info was just very useless for actually
finding out infos about the performance.
S
--=20
Sascha Runschke
Netzwerk- und Systemmanagement
Telefon : +49 (201) 102-1879 Mobil : +49 (173) 5419665 Fax : +49 (201)=20
102-1102105
GFKL Financial Services AG
Vorstand: Dr. Peter J=C3=A4nsch (Vors.), J=C3=BCrgen Baltes, Dr. Till Ergen=
zinger, Dr. Tom Haverkamp
Vorsitzender des Aufsichtsrats: Dr. Georg F. Thoma
Sitz: Limbecker Platz 1, 45127 Essen, Amtsgericht Essen, HRB 13522
--=_alternative 004A60DDC125741F_=
Content-Type: text/html; charset="utf-8"
content-transfer-encoding: quoted-printable
[email protected] schrieb
am 02.04.2008 14:55:15:
> While I agree this particular info is missing, the active check latenc=
y
> tells you how much you're lagging behind. As long as it's near 0 you'r=
e
> OK, but when it starts to show big latencies (i.e. >30 seconds)
even
> keeps growing then you definitely have a problem.
Yes I know. But the main point was, that the present=
ed
percentages
are more or less useless. The check latency is by
far the only real
relevant information on that page. A fact, that makes
the page seem
to be broken by design to me.
> Also note that service latency is not always a "server performanc=
e
> problem". In Nagios 2.0, host checks are run serially so the
time in
> takes to run host checks can cause latency to grow up pretty fast,
> especially when you start seeing hosts down.
>
> There's other reason for high latencies: perfdata commands, ocsd
> commands, etc. Ideally you can avoid this by having Nagios write a
file
> or FIFO instead... There's many example out there.
I do know, but that has actually nothing to do with
my initial proposal
This thread was not meant for debugging my performan=
ce
issues - especially
since I already solved them last week. ^^ It is more
or less just a reflection
of the fact, that the Performance Info was just very
useless for actually
finding out infos about the performance.
S
--
Sascha Runschk
...[email truncated]...
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]