Under the Metrics tab, add new metric that will be hidden in the chart and is used for alerting only. Duplicate the query and remove all template variables (i.e. $somevar) from it. Replace the template variable with a hard-coded value you want to create alert for. Hide the metric by clicking on the “eye” icon. … Read more
There are two ways that I know: You can use the $__interval variable like this: increase(http_requests_total[$__interval]) There is a drawback that the $__interval variable’s value is adjusted by resolution of the graph, but this may also be helpful in some situations. This approach should fit your case better: Go to Dashboard’s Templating settings, create new … Read more
Prometheus has an or logical binary operator, but what you’re asking about here is vector selectors. You can use a regex for this {job=~”traefik|cadvisor|prometheus”}, however that you want to do this is a smell.
It uses the query API endpoint rather than the query_range API endpoint on Prometheus, which is more efficient if you only care about the end of your time range and don’t want to pull in data that Grafana is going to throw away again.
New versions got a “Save As…” button in the dashboard menu:
count(count by (a) (hello_info)) First you want an aggregator with a result per value of a, and then you can count them.
My answer tries to elaborate on Carl’s answer. I assume that the GUI layout may have changed a little since 2016, so it took me while to find the “name” option. Assuming you have a metric as follows: # HELP db2_prometheus_adapter_info Information on the state of the DB2-Prometheus-Adapter # TYPE db2_prometheus_adapter_info gauge db2_prometheus_adapter_info{app_state=”UP”) 1.0 and … Read more
Thanks @Donald Mok for his answer; I just want to make it as clear as possible. In the Grafana interface you can create an organization. After that you can create some dashboards for this organization. So, there is a problem that you need to specify the organization for anonymous users. And it should be a … Read more