On Fri, May 29, 2020 at 9:13 AM Rajesh Reddy Nachireddi < [email protected]> wrote:
> Hi, > > 1. What is the cost of a alert rule in promethes? Interms of memory or cpu > cost- No.of timeseries matched or something more than that? > > Rules are queries, so it depends on the rule and how much data the rule evaluates. > > 2. How many rules are too many for a single prometheus ? > > Depends on the size of the server and the size, complexity, and frequency of your evaluation cycle. It could be single digits, it could be hundreds of thousands. 3. How large single alert rule can be ? > > As large as a single query can be executed. > 4. What is the cost of evaluation of a alert rule for which there are no > metrics available? Is it same as if the metric is available or negligible? > For example, if there is a rule on device=ip1 cpu metric, and there are no > device=ip1 metrics, what is the cost of that rule? > > Rules that match nothing are cheap. The index matching is reasonably efficient. > 5. What is best way to generate rules or relabel config for large enterprise > setups ? - Do we have some opensource/commercial tools which has UI or API > CRUD or Only through Config management tools ? > > We use code to generate much of our alerting configuration. Metrics per service are categorized and standardized to create SLO-driven data and alerts for all services. See: https://gitlab.com/gitlab-com/runbooks/-/tree/master/metrics-catalog > > Regards, > > Rajesh > > -- 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/CABbyFmofUOUHrxswmGZd_Mdb2BhoS%3D0cjSLRf9ShEv99uiA%2BXg%40mail.gmail.com.
