2:45 p.m.
It would be interesting if there were any official Prometheus exporter to monitor LDAP
metrics.
If we have one such exporter, does it authenticate LDAP Server to
aggregate/formulate/capture metrics?
Thanks,
Teja
3:42 p.m.
--On Wednesday, February 23, 2022 10:45 PM +0000 vtejaswini1(a)gmail.com
wrote:
It would be interesting if there were any official Prometheus
exporter to
monitor LDAP metrics. If we have one such exporter, does it authenticate
LDAP Server to aggregate/formulate/capture metrics?
I'm not familiar with Prometheus, but the monitor backend to OpenLDAP
provides a number of useful metrics that can be queried via standard ldap
operations. You can define a seperate set of ACL rules for that backend as
well.
This is an example of a dashboard I made for Grafana:
<https://grafana.com/grafana/dashboards/13349>
Other monitoring systems (for example datadog) also have plugins for
OpenLDAP that can be used to generate a similar dashboard.
Regards,
Quanah
2:46 a.m.
On 2/23/22 23:45, vtejaswini1(a)gmail.com wrote:
It would be interesting if there were any official Prometheus
exporter to monitor LDAP metrics.
What does "official" mean for you? Does that mean vendor support you're
willing to pay for?
Anyway...
My slapdcheck also produces OpenMetrics output to be consumed by
Prometheus. But since it's a wrapper around the MRPE checks it does not
feel Prometheus-like. It's on my long to-do list to refactor slapdcheck
to internally already handle labels and turn that into MRPE/Nagios
compatible check item names.
https://code.stroeder.com/ldap/slapdcheck
Help would be appreciated:
https://code.stroeder.com/ldap/slapdcheck/issues
You could also use mtail to extract metrics from logs which mtail also
serves as Prometheus metrics:
https://github.com/google/mtail
My mtail progs extract some metrics not available in cn=monitor yet
(e.g. result code counters and deferred counters).
https://code.stroeder.com/AE-DIR/ansible-ae-dir-server/src/branch/master/...
BTW: There are some simple exporters available out there which simply
pull counters and gauges from cn=monitor and use the DN of the monitor
entry as the metrics' label. Given that some cn=monitor entries contain
multiple values this is not the correct approach. (And full DNs as label
also makes your Prometheus queries look quite complicated.)
Ciao, Michael.
6 a.m.
Hello Michael,
Doesn't your slapdcheck also rely on cn=monitor to query LDAP Server? Could you
explain to me on a high level how slapdcheck is interacting with LDAP Server? I am happy
that the source code is open-sourced and available.
Coming from a Prometheus background, when I talk about official exporters I meant the
exporters listed by the Prometheus community as reliable and useful on their website. Here
is such a list: https://prometheus.io/docs/instrumenting/exporters/, we have a wide range
of exporters for different servers/databases.
Thanks,
Teja
7:28 a.m.
On 2/24/22 15:00, vtejaswini1(a)gmail.com wrote:
Doesn't your slapdcheck also rely on cn=monitor to query LDAP
Server?
Yes.
Could you explain to me on a high level how slapdcheck is
interacting
with LDAP Server?
Besides accessing cn=monitor it reads the currently active
configuration
from a read-only(!) cn=config and e.g. trys to reach all the provider
replicas configured therein, etc.
And in opposite to other slapd exporters it does not falsely assume to
only read one gauge/counter per entry in cn=monitor.
Coming from a Prometheus background, when I talk about official
exporters I meant the exporters listed by the Prometheus community as
reliable and useful on their website.
Ah, I see.
Not sure how much the Prometheus community care about OpenLDAP though.
Ciao, Michael.
401
days inactive
402
days old
openldap-technical@openldap.org
4 comments
3 participants
participants (3)
-
Michael Ströder -
Quanah Gibson-Mount -
vtejaswini1＠gmail.com