faas/prometheus/alert.rules.yml
Lucas Roesler 3bcc10a07e Upgrade prom and alert manager to 2.7.1 and 0.16.1
**What**
- Removes the `alert` label in the scale-up alert
- Updates the annotaitons to use the `function_name` label instead of
the `instance` label that was removed.
- Per prometheus/prometheus#4836 and the related mailing list discussion
https://groups.google.com/d/msg/prometheus-users/7Ul6ngc7Ogs/j_YDszV5BwAJ
the alert value should not be included in the alert labels otherwise
each calculation of the alert is treated like a new alert and then the
use of `for 5s` will not behave as expected.
- Ports the resoltuion openfaas/faas-netes#372

Signed-off-by: Lucas Roesler <roesler.lucas@gmail.com>
2019-02-21 09:51:38 +00:00

15 lines
444 B
YAML

groups:
- name: prometheus/alert.rules
rules:
- alert: service_down
expr: up == 0
- alert: APIHighInvocationRate
expr: sum(rate(gateway_function_invocation_total{code="200"}[10s])) BY (function_name) > 5
for: 5s
labels:
service: gateway
severity: major
annotations:
description: High invocation total on {{ $labels.function_name }}
summary: High invocation total on {{ $labels.function_name }}