On 25/01/2021 04:28, Rohit Bharati wrote:
Hi Everyone,
I am setting up monitoring for my bamboo Job(CI/CD tool). For which I
am using a nodeJS exporter ( I am new to it) which returns the
following feilds:
------------------------------------------------------------------------------------------------------------------------------------
*# HELP build_number build number for plan.*
*# TYPE build_number gauge*
*build_number{planName="Some_name"} 2839*
*
*
*# HELP successful_test_count number of test cases passed for the
build number.*
*# TYPE successful_test_count gauge*
*successful_test_count{planName="* *Some_name* *"} 3*
*
*
*# HELP failed_test_count number of test cases failed for the build
number.*
*# TYPE failed_test_count gauge*
*failed_test_count{planName="* *Some_name* *"} 0*
*
*
*# HELP test_count_percentage Percentage = (success/(success +
failure))*100 *
*# TYPE test_count_percentage gauge*
*test_count_percentage{planName="* *Some_name* *"} 100*
----------------------------------------------------------------------------------------------------------------------------------------------
as this plan is set to be built every 5 mins therefore I have set my
Prometheus to scrape this endpoint every 5 mins.
but in real case some times the plan takes more than 5 mins to build ,
hence the data updated after 5 mins so my Prometheus shows the
duplicate of the same metric in grafana
*What i wished*
I wanted to see results based on *build_number{planName="* *Some_name*
*"} 2839* which will be always unique and keep on increasing.
How can I achieve this so that I can see only unique values for a metric?
or how to make a metric value unique? either in promethues.yml or exporter
Prometheus should be set to scrape at least every 2 minutes, otherwise
time series will be marked as stale and not be returned by queries.
It is perfectly normal for metrics not to have changed between scrapes
for some applications, so nothing special needs to be done. Just set the
scrape interval to 2 minutes and let it keep fetching the latest values.
There are no worries around storage due to the way that the Prometheus
TSDB works.
Ideally you would use counters rather than gauges which count things
like successes in total (rather than for a particular run) to give
greater meaning & flexibility to the queries you can do. One thing to
remember is that Prometheus is a metrics system, not an events system,
so is designed to show how metrics change over time. There is no concept
of "a build" so it is not possible to find the number of successful
tests for a particular build, but it is possible to graph how that
number varies over time. If you do want to be able to lookup values
based on a build you should use an events system (such as Elasticsearch)
or a standard database (e.g. MySQL).
If you can describe what you are wanting to be able to display in
Grafana we might be able to give some more advice.
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/d9770407-f1bf-c47c-e19a-d22cb2e9a346%40Jahingo.com.
