Digging a bit more and this seems to be a local issue on my end.
Spun a grafana + prometheus on my laptop, used the same data and I see it
in Prometheus.
On Sun, Feb 27, 2022 at 12:43 PM Laurent Dumont
wrote:
> Forgot to add the runtime info!
>
> Runtime Information
> Start time
Build Information
Version 2.33.4
Revision 83032011a5d3e6102624fe58241a374a7201fee8
Branch HEAD
BuildUser root@d13bf69e7be8
BuildDate 20220222-16:51:28
GoVersion go1.17.7
On Sun, Feb 27, 2022 at 12:40 PM Laurent Dumont
wrote:
> Hey folks,
>
> Trying to load some past data into Prometh
Hey folks,
Trying to load some past data into Prometheus using create-blocks-from.
- The data appears to be proper Openmetrics formatted.
- No errors from Prometheus
- I do see the metric in Prometheus, but there doesnt appear to be any
data points.
- Tried restarting the server to
Hi!
Trying to wrap my brain around the group_left / group_right in PromQL.
I have the following metrics (removed some labels)
potato{short_hostname="qasite1-compute001",site="qasite1",hardware="hardwaretype1"}
1
potato{short_hostname="qasite1-compute001",site="qasite1",hardware="hardwaretype2"}
I don't know how GCP calculates their CPU metrics, but
node_cpu_seconds_total looks to contain statistics for
user/kernel/interrupt etc spaces. Maybe you can make a separate graph based
on each of those and see if one is much higher (
https://www.robustperception.io/understanding-machine-cpu-usage)
I dont think that will work out of the box. Prometheus doesn't understand
Elasticsearch by itself.
You could use this https://github.com/braedon/prometheus-es-exporter to
export the data inside ES to a Prometheus standard format. You can then
scrape that target from your Prometheus server and aler
Salut!
First thing would be to look at the actual service logs of the Prometheus
service.
Something like :
journalctl -u prometheus.service
That should tell you what is failing to start and forcing systemd to kill
the service.
On Tue, May 4, 2021 at 10:38 AM Support IT Adept wrote:
> Hi,
>
This is a general help ML with no expectations of support.
You will need to provide more details regarding how you are getting metrics
for CPU usage and if it's available on a per-core basis.
On Tue, Mar 23, 2021 at 4:16 PM Vamshi Bhargav
wrote:
> Any update on this.
>
> On Tuesday, March 23, 2
What docker-compose command are you using? You need to use "docker-compose
up" to actually build + start the container.
On Tue, Mar 23, 2021 at 8:30 PM Anauê Curi wrote:
> Hi everyone,
>
> I'm trying to run prometheus via a container in docker, but after run it
> the container status moves to cr
I don't think your use case is something that AM or Prometheus is looking
to solve.
The way I see it :
- Prometheus has metrics and alarm patterns.
- It triggers an alarm and sends it to AM.
- AM receives the alarm and does some basic routing based on labels.
- Once the Prometheus pat
This is usually something solved outside of Prometheus.
Prometheus container logs --> Fluentd --> ElasticSearch --> Kibana display
You can then use Kibana to change the display of the time to your local
timezone. A pretty common design decision is to log everything in UTC and
let other systems pa
If you are coming from a Zabbix world, there is no equivalent maintenance
mechanic in Prometheus. Most of the service assurance/service mapping is
not part of the Prometheus stack.
Silences are only active between X and Y and will expire after Y. They
cannot repeat so you'll need some kind of exte
It depends on what you want to test
Actual Prometheus performance
-
https://blog.freshtracks.io/load-testing-prometheus-metric-ingestion-5b878711711c
- https://github.com/thanos-io/thanosbench
I'm not sure just testing a GET on prom-ip:9090/ is going to tell you much.
Maybe a bunch of
Thanks!
I missed that part. I'll tinker around with it.
On Thu, Nov 19, 2020 at 5:55 PM Christian Hoffmann <
m...@hoffmann-christian.info> wrote:
> Hi,
>
> On 11/19/20 11:32 PM, Laurent Dumont wrote:
> > collectd_openstack_nova_gauge{exported_instance="
> s
Hi!
I was trying to use metric_relabel_configs in order to change the label
content (sourcing from another label) of a metric.
The original metric looks like this:
collectd_openstack_nova_gauge{exported_instance="site1-director.potato.com
",instance="123.123.123.123:9103",job="collectors",openst
I don't think there would be any issues with multiple jobs with one target
OR one job with multiple targets.
Is the data being exposed by each PG different?
On Sun, Nov 15, 2020 at 9:57 PM Andy Pan wrote:
> Is there a right path forward for Prometheus to collect metrics from
> multiple push gat
You can add some HTTP auth in front of the PG (not directly though PG but
with nginx/apache). But yes, as long as the pod needs to access PG
directly, I guess it means that the customer running code inside the pod
will be able to talk to PG.
On Thu, Nov 12, 2020 at 10:19 AM Rafael Paulovic
wrote:
I'm not sure if I fully understand the flow of metrics, but you can use a
PushGateway as a "central" scrape target. In your case, since your
application pods seem to be short-lived, a per pod scrape architecture
might not be great. You can use the prometheus_client library to push
application metri
If the blackbox exporter itself doesn't seem to respond/load properly, I
would troubleshoot that first. You can look at the pod logs.
On Thu, Nov 12, 2020 at 2:36 AM nesa...@gmail.com
wrote:
> I have a blackbox exporter deployed on a K8S cluster, The problem is, some
> URLs (which are using the
ance
>
> On Mon, Nov 9, 2020 at 10:06 Laurent Dumont
> wrote:
>
>> Your retention is 16 days and you filled the data volume in 13?
>>
>> I don't Prometheus will alarm for it's data partition by default.
>>
>> On Mon, Nov 9, 2020, 12:27 PM Laurent De
How many nodes do you have as targets? Are all of the 2000 rules the same?
With a large number of nodes and a rule that is returning a large number of
metrics, it might be a lot to process.
On Tue, Nov 10, 2020 at 1:18 AM 万锐 wrote:
> On Tuesday, November 10, 2020 at 2:16:02 PM UTC+8 万锐 wrote:
>
Your retention is 16 days and you filled the data volume in 13?
I don't Prometheus will alarm for it's data partition by default.
On Mon, Nov 9, 2020, 12:27 PM Laurent Demailly wrote:
> *What did you do?*
> Setup using community helm chart (8Gi data volume, 15 days retention)
>
> *What did you
; Thanks for your reply!
>
> I agree, but the the state of the art continues to be that this ...
> doesn't work very well. Even RedHat's KVM+QEMU documentation recommends
> that the guests run NTP.
>
> I'm still curious what the source of the issue is. Laurent Dumon
When you resume the VM, does NTP resync? I guess that since you just pause
the VM, once it unpause, the time will be incorrect and that would explain
the out-of-order message. The scaped host will be in the future from the
perspective of the Prometheus server.
On Wed, Nov 4, 2020 at 8:46 PM Harald
I wonder if you could use a Push Gateway to which all the spotty hosts
would write to. They could write locally the metrics to a file and keep
trying to push the metrics once they have the connectivity to the central
Prometheus.
On Sat, Aug 22, 2020 at 1:43 PM Brian Candler wrote:
> Another opti
You could use a templated tile. You have a variable that represents every
system monitored by Prometheus and the same tile is replicated for every
value of that variable.
On Fri, Aug 14, 2020, 6:22 PM Thomas Berger
wrote:
> For Info:
>
> I use a Dashboard in Grafana with Tiles.
> Each Tile stand
I don't think Prometheus can do much in that case.
It's end goal is to expose a metric. With Alert rules, you can tie some
level of business logic. What we do is standardize our exporting values,
where 1 = OK and 0 = NOT OK for custom data. You can then have alert rules
that will alarm based on th
I think you need to look at the source code for the Openstack exporter :)
On Fri, Aug 14, 2020 at 7:49 AM Saurabh Jain
wrote:
> Hi Everyone ...
>
>
> I need your advice related to openstack_neutron_up metric from openstack
> exporter.The value which we are getting for this metric is 0 .What does
That something that I investigated as well. So far, it seems a silence will
mute an alarm at the root level on the AM side. There is no notifications
that a silence was created either.
In your case, you could have the reverse where your monitoring system could
create the silences locally and push
Hi!
We have a pretty specific use case where we have three different recipients
for Alerts.
-Recipient1
-Recipient2
-Recipient3
Right now, if I create a silence, the Alarms matching that silence will be
ignored for all three Recipients.
I was wondering if there was a way to create a silence b
30 matches
Mail list logo