fix(charts/istio-alerts): Remove source service from 5XX Alarm

charts/istio-alerts/Chart.yaml

@@ -2,7 +2,7 @@ apiVersion: v2
 name: istio-alerts
 description: A Helm chart that provisions a series of alerts for istio VirtualServices
 type: application
-version: 0.3.2
+version: 0.4.0
 appVersion: 0.0.1
 maintainers:
   - name: diranged

charts/istio-alerts/README.md

@@ -1,6 +1,6 @@
 # istio-alerts
 
-![Version: 0.3.2](https://img.shields.io/badge/Version-0.3.2-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 0.0.1](https://img.shields.io/badge/AppVersion-0.0.1-informational?style=flat-square)
+![Version: 0.4.0](https://img.shields.io/badge/Version-0.4.0-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 0.0.1](https://img.shields.io/badge/AppVersion-0.0.1-informational?style=flat-square)
 
 A Helm chart that provisions a series of alerts for istio VirtualServices
 
@@ -13,6 +13,15 @@ A Helm chart that provisions a series of alerts for istio VirtualServices
 
 ## Upgrade Notes
 
+### 0.3.x -> 0.4.x
+
+**BREAKING: http5XXMonitor no longer alerts per source client workload.**
+
+In version 0.2.x, there was a change to the default `http5XXMonitor` which
+introduced calculation of the error rate per source workload. This 0.4.x
+release reverts this behavior by default while allowing you to opt in to custom
+selectors via the `monitorGroupingLabels` option.
+
 ### 0.2.x -> 0.3.x
 
 **BREAKING: The DestinationServiceSelectorValidity alert rule requires kube-state-metrics.**
@@ -22,6 +31,12 @@ you do not have kube-state-metrics installed, you will need to disable the alert
 `serviceRules.destinationServiceSelectorValidity.enabled` to `false`. This alert is used to detect
 if the destinationServiceSelector is actually selecting series for a service that exists.
 
+### 0.2.x
+
+**BREAKING: http5XXMonitor now calculcates the 5XX error rate for each client**
+source workload using the `source_workload` label, and will alert if any
+`source_workload`'s error rate exceeds the specified `threshold`.
+
 ## Values
 
 | Key | Type | Default | Description |
@@ -48,10 +63,10 @@ if the destinationServiceSelector is actually selecting series for a service tha
 | serviceRules.highRequestLatency.percentile | float | `0.95` | Which percentile to monitor - should be between 0 and 1. Default is 95th percentile. |
 | serviceRules.highRequestLatency.severity | string | `"warning"` | Severity of the latency monitor |
 | serviceRules.highRequestLatency.threshold | float | `0.5` | The threshold for considering the latency monitor to be alarming. This is in seconds. |
-| serviceRules.http5XXMonitor | object | `{"enabled":true,"for":"5m","monitorGroupingLabels":["destination_service_name","reporter","source_workload"],"severity":"critical","threshold":0.0005}` | Configuration related to the 5xx monitor for the VirtualService. |
+| serviceRules.http5XXMonitor | object | `{"enabled":true,"for":"5m","monitorGroupingLabels":["destination_service_name","reporter"],"severity":"critical","threshold":0.0005}` | Configuration related to the 5xx monitor for the VirtualService. |
 | serviceRules.http5XXMonitor.enabled | bool | `true` | Whether to enable the monitor on 5xxs returned by the VirtualService. |
 | serviceRules.http5XXMonitor.for | string | `"5m"` | How long to evaluate the rate of 5xxs over. |
-| serviceRules.http5XXMonitor.monitorGroupingLabels | list | `["destination_service_name","reporter","source_workload"]` | The set of labels to use when evaluating the ratio of the 5XX. |
+| serviceRules.http5XXMonitor.monitorGroupingLabels[0] | string | `"destination_service_name"` | The set of labels to use when evaluating the ratio of the 5XX. |
 | serviceRules.http5XXMonitor.severity | string | `"critical"` | Severity of the 5xx monitor |
 | serviceRules.http5XXMonitor.threshold | float | `0.0005` | The threshold for considering the 5xx monitor to be alarming. Default is 0.05% error rate, i.e 99.95% reliability. |
 

charts/istio-alerts/README.md.gotmpl

@@ -8,6 +8,15 @@
 
 ## Upgrade Notes
 
+### 0.3.x -> 0.4.x
+
+**BREAKING: http5XXMonitor no longer alerts per source client workload.**
+
+In version 0.2.x, there was a change to the default `http5XXMonitor` which
+introduced calculation of the error rate per source workload. This 0.4.x
+release reverts this behavior by default while allowing you to opt in to custom
+selectors via the `monitorGroupingLabels` option.
+
 ### 0.2.x -> 0.3.x
 
 **BREAKING: The DestinationServiceSelectorValidity alert rule requires kube-state-metrics.**
@@ -17,6 +26,12 @@ you do not have kube-state-metrics installed, you will need to disable the alert
 `serviceRules.destinationServiceSelectorValidity.enabled` to `false`. This alert is used to detect
 if the destinationServiceSelector is actually selecting series for a service that exists.
 
+### 0.2.x
+
+**BREAKING: http5XXMonitor now calculcates the 5XX error rate for each client**
+source workload using the `source_workload` label, and will alert if any
+`source_workload`'s error rate exceeds the specified `threshold`.
+
 {{ template "chart.requirementsSection" . }}
 
 {{ template "chart.valuesSection" . }}

charts/istio-alerts/values.yaml

@@ -99,11 +99,10 @@ serviceRules:
     # -- Severity of the 5xx monitor
     severity: critical
 
-    # -- The set of labels to use when evaluating the ratio of the 5XX.
     monitorGroupingLabels:
+      # -- The set of labels to use when evaluating the ratio of the 5XX.
       - destination_service_name
       - reporter
-      - source_workload
 
   # -- Configuration related to the latency monitor for the VirtualService.
   highRequestLatency: