SLO Dashboard

Tracks service's SLO and error budget consumption.

Overview
Revisions
Reviews
Edit →

The dashboard is based on the following example: https://github.com/google/prometheus-slo-burn-example

The dashboard assumes that there is a probe, which probes service’s availability and exports the data using metrics, which are then used by this dashboard. There are supposed to be two metrics exposed, which can have any name (it is configurable in the parameters):

requests_error_count - the number of requests to the service returned an error;
requests_total_count - the total number of requests to the service.

Settings for color switching are needed to be adjusted:

After error_budget parameter is adjusted for the required level, it is necessary to adjust the thresholds for color switching at Error Budget Left singlestat panel. It can be done via Vizualization tab.
After service_slo_target parameter is adjusted for the required level, it is necessary to adjust the thresholds for color switching at Current SLO singlestat panel. It can be done via Vizualization tab.

The highest panel has two states:

Status is OK - there are no alerts firing for the last 5 minutes.
Alerts are firing - there are alerts firing for the last 5 minutes, pay attention.

The titles for both states can be changed via Vizualization tab, in Value mappings section.

Welcome to contribute!