I just upgraded one of our servers to dovecot v2.1.16 (ee), and am looking into the stats feature. Am I interpreting the wiki correct in reading that the "doveadm stats dump command" only returns statistics about IMAP commands?
Is it possible to collect info about POP3 and LMTP commands also ?
Also, is "doveadm stats dump command" telling me the results of all commands that has finished the last stats_command_min_time, or will it maybe contain much more than 1 minute of activity ?
-jf
On 29.4.2013, at 15.30, Jan-Frode Myklebust <janfrode@tanso.net> wrote:
I just upgraded one of our servers to dovecot v2.1.16 (ee), and am looking into the stats feature. Am I interpreting the wiki correct in reading that the "doveadm stats dump command" only returns statistics about IMAP commands?
Right.
Is it possible to collect info about POP3 and LMTP commands also ?
No. I think they would be pretty boring statistics, since with POP3 pretty much everything causing disk I/O or CPU usage would be RETRs and with LMTP everything would be DATAs.
Also, is "doveadm stats dump command" telling me the results of all commands that has finished the last stats_command_min_time, or will it maybe contain much more than 1 minute of activity ?
It can contain much more. The stats process will keep as much data in memory as possible until it reaches the stats_memory_limit. The doveadm stats dump lists everything that the stats process knows.
On Mon, May 6, 2013 at 4:37 PM, Timo Sirainen <tss@iki.fi> wrote:
On 29.4.2013, at 15.30, Jan-Frode Myklebust <janfrode@tanso.net> wrote:
Is it possible to collect info about POP3 and LMTP commands also ?
No. I think they would be pretty boring statistics, since with POP3 pretty much everything causing disk I/O or CPU usage would be RETRs and with LMTP everything would be DATAs.
I think knowing the timings of writing messages to disk / reading from disk would be very interesting and relevant data. Especially for us with mostly POP3 clients, where LMTP DATAs and POP3 RETRs probably is accounting for major parts of the server load.
Also, is "doveadm stats dump command" telling me the results of all commands that has finished the last stats_command_min_time, or will it maybe contain much more than 1 minute of activity ?
It can contain much more. The stats process will keep as much data in memory as possible until it reaches the stats_memory_limit. The doveadm stats dump lists everything that the stats process knows.
Ok, then I guess we'll need to limit our stats dumps based on last_seen.
-jf
Am 08.05.2013 10:25, schrieb Jan-Frode Myklebust:
On Mon, May 6, 2013 at 4:37 PM, Timo Sirainen <tss@iki.fi> wrote:
On 29.4.2013, at 15.30, Jan-Frode Myklebust <janfrode@tanso.net> wrote:
Is it possible to collect info about POP3 and LMTP commands also ?
No. I think they would be pretty boring statistics, since with POP3 pretty much everything causing disk I/O or CPU usage would be RETRs and with LMTP everything would be DATAs.
I think knowing the timings of writing messages to disk / reading from disk would be very interesting and relevant data. Especially for us with mostly POP3 clients, where LMTP DATAs and POP3 RETRs probably is accounting for major parts of the server load.
no urgent need for stats i count pop3 logins with xymon out of syslog in realtime, also logwatch reports lmtp/pop3/imap daily for each user by cron in daily terms
perhaps have a look at
http://sys4.de/de/blog/2013/01/10/xymon-dovecot-count-imap-pop3-logins-graph...
if you have xymon monitoring etc installed you have other stuff like cpu , mem, disk io too
i ll plan to write some solution for xymon retrieve data out of dovecot stats too
Also, is "doveadm stats dump command" telling me the results of all commands that has finished the last stats_command_min_time, or will it maybe contain much more than 1 minute of activity ?
It can contain much more. The stats process will keep as much data in memory as possible until it reaches the stats_memory_limit. The doveadm stats dump lists everything that the stats process knows.
Ok, then I guess we'll need to limit our stats dumps based on last_seen.
-jf
Best Regards MfG Robert Schetterer
-- [*] sys4 AG
http://sys4.de, +49 (89) 30 90 46 64 Franziskanerstraße 15, 81669 München
Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263 Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer Aufsichtsratsvorsitzender: Florian Kirstein
Thanks, nice graphs. I've attached a graph over LMTP delays per minute as seen from the postfix side on one of our servers. This includes delays caused by both delivery to dovecot LMTP, and also LMTP communication internally on the mailservers between postfix and amavis. Unfortunately it says nothing about the delivery time to each individual dovecot backend, since these are hiding behind dovecot director, and therefor we have no way of knowing which of our backends are slow (if any).
-jf
Am 08.05.2013 11:25, schrieb Jan-Frode Myklebust:
Thanks, nice graphs. I've attached a graph over LMTP delays per minute as seen from the postfix side on one of our servers. This includes delays caused by both delivery to dovecot LMTP, and also LMTP communication internally on the mailservers between postfix and amavis. Unfortunately it says nothing about the delivery time to each individual dovecot backend, since these are hiding behind dovecot director, and therefor we have no way of knowing which of our backends are slow (if any).
-jf
i am not sure, i ll have to read more doku perhaps it makes sense to use delay stuff in log for analyse i.e after all i am not using directors, i have "real" loadbalancers
private/dovecot-lmtp], delay=1.4, delays=1.2/0/0/0.17, dsn=2.0.0, status=sent
Best Regards MfG Robert Schetterer
-- [*] sys4 AG
http://sys4.de, +49 (89) 30 90 46 64 Franziskanerstraße 15, 81669 München
Sitz der Gesellschaft: München, Amtsgericht München: HRB 199263 Vorstand: Patrick Ben Koetter, Axel von der Ohe, Marc Schiffbauer Aufsichtsratsvorsitzender: Florian Kirstein
participants (3)
-
Jan-Frode Myklebust
-
Robert Schetterer
-
Timo Sirainen