On Thursday, June 18, 2020 at 2:49:35 PM UTC+2, Ben Kochie wrote: > > > > On Thu, Jun 18, 2020 at 2:41 PM Juergen Etzlstorfer < > [email protected] <javascript:>> wrote: > >> That's interesting! >> Which metrics do you usually focus on? >> > > We're following a variation of the RED method. > * Apdex scores for duration. > * Request rates > * Error rates > * Component saturation > > >> Could you please provide a bit more context on how you are then using the >> jsonnet files? >> > > The jsonnet files allow us to define what metrics are associated with what > component. For example, "these histogram buckets are used for apdex". > > This allows us to generate standardized recording rules, as well as > populate automated Grafana dashboards. > Are you also using the files for automated generation of alerts?
> > >> When is the evaluation triggered? Each time a new microservice gets >> deployed or by manual execution? >> > > The jsonnet files are in the repo, and get built and published in CI > pipelines. Whenever we change our services, if that's sharding changes, new > services, or new metrics for services. > That sounds similar to the approach we are taking. > > >> >> >> >> On Thursday, June 18, 2020 at 2:34:24 PM UTC+2, Ben Kochie wrote: >>> >>> We're doing something very similar. We have a metrics catalog[0] that >>> turns all of our service level metrics into uniform SLO metrics. >>> >>> [0]: >>> https://gitlab.com/gitlab-com/runbooks/-/tree/master/metrics-catalog >>> >>> On Thu, Jun 18, 2020 at 2:25 PM Juergen Etzlstorfer < >>> [email protected]> wrote: >>> >>>> Hi, >>>> >>>> I am a core contributor to an open-source project which makes use of >>>> Prometheus data for the evaluation of quality criteria of a microservice >>>> or >>>> an application. Basically, we automatically pull the different metrics and >>>> evaluate them against a given specification file based on service-level >>>> objectives (SLOs). We usually use SLOs based on the RED metrics such as >>>> throughput, error rate, and duration/response time. >>>> We want to provide more integrations out-of-the-box and I would be >>>> interested in which metrics would you usually like to check to determine >>>> the quality of your software? >>>> >>>> If you want to take a look at how we are doing it, we just recently >>>> recorded a webinar demonstrating this use case. >>>> https://www.youtube.com/watch?v=hx0NHj4u7ic >>>> >>>> The project itself can be found here if you are interested: >>>> https://keptn.sh >>>> >>>> Looking forward to hearing your thoughts on this! >>>> >>>> >>>> -- >>>> 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/2d816509-19d4-4969-9a77-7b10b9020427o%40googlegroups.com >>>> >>>> <https://groups.google.com/d/msgid/prometheus-users/2d816509-19d4-4969-9a77-7b10b9020427o%40googlegroups.com?utm_medium=email&utm_source=footer> >>>> . >>>> >>> -- >> 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] <javascript:>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/prometheus-users/7864ef9f-06d4-4542-b3fa-2fa257adaf14o%40googlegroups.com >> >> <https://groups.google.com/d/msgid/prometheus-users/7864ef9f-06d4-4542-b3fa-2fa257adaf14o%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- 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/c4330d25-70de-4e92-83db-434977037ec4o%40googlegroups.com.
