95 lines
4.9 KiB
Markdown
95 lines
4.9 KiB
Markdown
# CAAS KUBERNETES INGRESS VTS DataDog monitors
|
|
|
|
## How to use this module
|
|
|
|
```
|
|
module "datadog-monitors-caas-kubernetes-ingress-vts" {
|
|
source = "claranet/monitors/datadog//caas/kubernetes/ingress/vts"
|
|
version = "{revision}"
|
|
|
|
environment = var.environment
|
|
message = module.datadog-message-alerting.alerting-message
|
|
}
|
|
|
|
```
|
|
|
|
## Purpose
|
|
|
|
Creates DataDog monitors with the following checks:
|
|
|
|
- Nginx Ingress 4xx errors
|
|
- Nginx Ingress 5xx errors
|
|
|
|
## Inputs
|
|
|
|
| Name | Description | Type | Default | Required |
|
|
|------|-------------|------|---------|:-----:|
|
|
| artificial\_requests\_count | Number of false requests used to mitigate false positive in case of low trafic | `number` | `5` | no |
|
|
| environment | Architecture Environment | `string` | n/a | yes |
|
|
| evaluation\_delay | Delay in seconds for the metric evaluation | `number` | `15` | no |
|
|
| filter\_tags\_custom | Tags used for custom filtering when filter\_tags\_use\_defaults is false | `string` | `"*"` | no |
|
|
| filter\_tags\_custom\_excluded | Tags excluded for custom filtering when filter\_tags\_use\_defaults is false | `string` | `""` | no |
|
|
| filter\_tags\_use\_defaults | Use default filter tags convention | `string` | `"true"` | no |
|
|
| ingress\_4xx\_enabled | Flag to enable Ingress 4xx errors monitor | `string` | `"true"` | no |
|
|
| ingress\_4xx\_extra\_tags | Extra tags for Ingress 4xx errors monitor | `list(string)` | `[]` | no |
|
|
| ingress\_4xx\_message | Message sent when an alert is triggered | `string` | `""` | no |
|
|
| ingress\_4xx\_threshold\_critical | 4xx critical threshold in percentage | `string` | `"40"` | no |
|
|
| ingress\_4xx\_threshold\_warning | 4xx warning threshold in percentage | `string` | `"20"` | no |
|
|
| ingress\_4xx\_time\_aggregator | Monitor aggregator for Ingress 4xx errors [available values: min, max or avg] | `string` | `"min"` | no |
|
|
| ingress\_4xx\_timeframe | Monitor timeframe for Ingress 4xx errors [available values: `last_#m` (1, 5, 10, 15, or 30), `last_#h` (1, 2, or 4), or `last_1d`] | `string` | `"last_5m"` | no |
|
|
| ingress\_5xx\_enabled | Flag to enable Ingress 5xx errors monitor | `string` | `"true"` | no |
|
|
| ingress\_5xx\_extra\_tags | Extra tags for Ingress 5xx errors monitor | `list(string)` | `[]` | no |
|
|
| ingress\_5xx\_message | Message sent when an alert is triggered | `string` | `""` | no |
|
|
| ingress\_5xx\_threshold\_critical | 5xx critical threshold in percentage | `string` | `"20"` | no |
|
|
| ingress\_5xx\_threshold\_warning | 5xx warning threshold in percentage | `string` | `"10"` | no |
|
|
| ingress\_5xx\_time\_aggregator | Monitor aggregator for Ingress 5xx errors [available values: min, max or avg] | `string` | `"min"` | no |
|
|
| ingress\_5xx\_timeframe | Monitor timeframe for Ingress 5xx errors [available values: `last_#m` (1, 5, 10, 15, or 30), `last_#h` (1, 2, or 4), or `last_1d`] | `string` | `"last_5m"` | no |
|
|
| message | Message sent when an alert is triggered | `any` | n/a | yes |
|
|
| new\_host\_delay | Delay in seconds before monitor new resource | `number` | `300` | no |
|
|
| notify\_no\_data | Will raise no data alert if set to true | `bool` | `true` | no |
|
|
| prefix\_slug | Prefix string to prepend between brackets on every monitors names | `string` | `""` | no |
|
|
|
|
## Outputs
|
|
|
|
| Name | Description |
|
|
|------|-------------|
|
|
| nginx\_ingress\_too\_many\_4xx\_id | id for monitor nginx\_ingress\_too\_many\_4xx |
|
|
| nginx\_ingress\_too\_many\_5xx\_id | id for monitor nginx\_ingress\_too\_many\_5xx |
|
|
|
|
## Related documentation
|
|
|
|
DataDog blog: https://www.datadoghq.com/blog/monitor-prometheus-metrics
|
|
https://github.com/kubernetes/ingress-nginx/pull/423/commits/1d38e3a38425f08de2f75fcae13896a3fec4d144
|
|
|
|
## Nginx Ingress Controller setup
|
|
|
|
This configuration and monitors only work for ingress controller version :
|
|
- \>= 0.10 because ingress is beta before that and metrics naming convention not stable
|
|
- <= 0.15 because ingress does not use VTS metrics since 0.16
|
|
Enable the following flags in the Nginx Ingress Controller chart
|
|
controller.stats.enabled=true,controller.metrics.enabled=true
|
|
and the following Datadog agent configuration for each ingress controller:
|
|
```
|
|
datadog:
|
|
confd:
|
|
prometheus.yaml: |-
|
|
#nginx_upstream_responses_total{ingress_class,namespace,server,status_code:{1xx,2xx,3xx,4xx,5xx},upstream}
|
|
#nginx_upstream_requests_total{ingress_class,namespace,server,upstream}
|
|
init_config:
|
|
instances:
|
|
# The prometheus endpoint to query from
|
|
- prometheus_url: http://nginx-ingress-controller-metrics:9913/metrics
|
|
# This is NOT the ingress namespace, it is the prefix that will be used for the custom metrics
|
|
namespace: nginx-ingress
|
|
# Filter on the following metrics only
|
|
metrics:
|
|
- "nginx_upstream_requests_total"
|
|
- "nginx_upstream_responses_total"
|
|
# Adapt the tags to the current convention and verify that the monitor will match
|
|
tags:
|
|
- dd_monitoring:enabled
|
|
- dd_ingress:enabled
|
|
- dd_ingress_class:nginx
|
|
- env:ENV
|
|
```
|