Add CoderdUnprovisionedPrebuiltWorkspaces alert

README.md

@@ -244,7 +244,7 @@ values which are defined [here](https://github.com/grafana/helm-charts/tree/main
 
 | Key | Type | Default | Description |
 |-----|------|---------|-------------|
-| global.coder.alerts | object | `{"coderd":{"groups":{"CPU":{"delay":"10m","enabled":true,"period":"10m","thresholds":{"critical":0.9,"warning":0.8}},"IneligiblePrebuilds":{"delay":"10m","enabled":true,"thresholds":{"notify":1}},"Memory":{"delay":"10m","enabled":true,"thresholds":{"critical":0.9,"warning":0.8}},"Replicas":{"delay":"5m","enabled":true,"thresholds":{"critical":1,"notify":3,"warning":2}},"Restarts":{"delay":"1m","enabled":true,"period":"10m","thresholds":{"critical":3,"notify":1,"warning":2}},"WorkspaceBuildFailures":{"delay":"10m","enabled":true,"period":"10m","thresholds":{"critical":10,"notify":2,"warning":5}}}},"enterprise":{"groups":{"Licences":{"delay":"1m","enabled":true,"thresholds":{"critical":1,"warning":0.9}}}},"provisionerd":{"groups":{"Replicas":{"delay":"5m","enabled":true,"thresholds":{"critical":1,"notify":3,"warning":2}}}}}` | alerts for the various aspects of Coder |
+| global.coder.alerts | object | `{"coderd":{"groups":{"CPU":{"delay":"10m","enabled":true,"period":"10m","thresholds":{"critical":0.9,"warning":0.8}},"IneligiblePrebuilds":{"delay":"10m","enabled":true,"thresholds":{"notify":1}},"Memory":{"delay":"10m","enabled":true,"thresholds":{"critical":0.9,"warning":0.8}},"Replicas":{"delay":"5m","enabled":true,"thresholds":{"critical":1,"notify":3,"warning":2}},"Restarts":{"delay":"1m","enabled":true,"period":"10m","thresholds":{"critical":3,"notify":1,"warning":2}},"UnprovisionedPrebuiltWorkspaces":{"delay":"10m","enabled":true,"thresholds":{"warn":1}},"WorkspaceBuildFailures":{"delay":"10m","enabled":true,"period":"10m","thresholds":{"critical":10,"notify":2,"warning":5}}}},"enterprise":{"groups":{"Licences":{"delay":"1m","enabled":true,"thresholds":{"critical":1,"warning":0.9}}}},"provisionerd":{"groups":{"Replicas":{"delay":"5m","enabled":true,"thresholds":{"critical":1,"notify":3,"warning":2}}}}}` | alerts for the various aspects of Coder |
 | global.coder.coderdSelector | string | `"pod=~`coder.*`, pod!~`.*provisioner.*`"` | series selector for Prometheus/Loki to locate provisioner pods. ensure this uses backticks for quotes! |
 | global.coder.controlPlaneNamespace | string | `"coder"` | the namespace into which the control plane has been deployed. |
 | global.coder.externalProvisionersNamespace | string | `"coder"` | the namespace into which any external provisioners have been deployed. |

coder-observability/runbooks/coderd.md

@@ -82,4 +82,54 @@ Please contact your Coder sales contact, or visit https://coder.com/contact/sale
 Prebuilds only become eligible to be claimed by users once the workspace's agent is a) running and b) all of its startup
 scripts have completed.
 
-If a prebuilt workspace is not eligible, view its agent logs to diagnose the problem.
+If a prebuilt workspace is not eligible, view its agent logs to diagnose the problem.
+
+## CoderdUnprovisionedPrebuiltWorkspaces
+
+The number of running prebuilt workspaces is lower than the desired instances. This could be for several reasons,
+ordered by likehood:
+
+### Experiment/License
+
+The prebuilds feature is currently gated behind an experiment *and* a premium license.
+
+Ensure that the prebuilds experiment is enabled with `CODER_EXPERIMENTS=workspace-prebuilds`, and that you have a premium
+license added.
+
+### Preset Validation Issue
+
+Templates which have prebuilds configured will require a configured preset defined, with ALL of the required parameters
+set in the preset. If any of these are missing, or any of the parameters - as defined - fail validation, then the prebuilds
+subsystem will refuse to attempt a workspace build.
+
+Consult the coderd logs for more information; look out for errors or warnings from the prebuilds subsystem.
+
+### Template Misconfiguration or Error
+
+Prebuilt workspaces cannot be provisioned due to some issue at `terraform apply`-time. This could be due to misconfigured
+cloud resources, improper authorization, or any number of other issues.
+
+Visit the Workspaces page, change the search term to `owner:prebuilds`, and view on the previously failed builds. The
+error will likely be quite obvious.
+
+### Provisioner Latency
+
+If your provisioners are overloaded and cannot process provisioner jobs quickly enough, prebuilt workspaces may be affected.
+There is no prioritization at present for prebuilt workspace jobs.
+
+Ensure your provisioners are appropriately resources (i.e. you have enough instances) to handle the concurrent build demand.
+
+### Use of Workspace Tags
+
+If you are using `coder_workspace_tags` ([docs](https://coder.com/docs/admin/templates/extending-templates/workspace-tags))
+in your template, chances are you do not have any provisioners running or they are under-resourced (see **Provisioner Latency**).
+
+Ensure your running provisioners are configured with your desired tags.
+
+### Reconciliation Loop Issue
+
+The prebuilds subsystem runs a _reconciliation loop_ which monitors the state of prebuilt workspaces to ensure the desired
+number of instances are present at all times. Workspace Prebuilds is currently a BETA feature and so there could be a bug
+in this _reconciliation loop_, which should be reported to Coder.
+
+Examine your coderd logs for any errors or warnings relating to prebuilds.

coder-observability/templates/configmap-prometheus-alerts.yaml

@@ -125,6 +125,26 @@ data:
   {{- end }}
   {{- end }}
 
+  {{- with .groups.UnprovisionedPrebuiltWorkspaces }}
+  {{- $group := . }}
+  {{- if .enabled }}
+    - name: Coderd Unprovisioned Prebuilt Workspaces
+      rules:
+      {{ $alert := "CoderdUnprovisionedPrebuiltWorkspaces" }}
+      {{- range $severity, $threshold := .thresholds }}
+      - alert: {{ $alert }}
+        expr: max by (template_name, preset_name) (coderd_prebuilds_desired - coderd_prebuilds_running) > 0
+        for: {{ $group.delay }}
+        annotations:
+          summary: >
+            {{ `{{ $value }}` }} prebuilt workspace(s) not yet been provisioned for the "{{ `{{ $labels.template_name }}` }}" template and "{{ `{{ $labels.preset_name }}` }}" preset.
+        labels:
+          severity: {{ $severity }}
+          runbook_url: {{ template "runbook-url" (deepCopy $ | merge (dict "alert" $alert) $service) }}
+      {{- end }}
+  {{- end }}
+  {{- end }}
+
   {{- end }} {{/* end-section */}}
 
 

coder-observability/templates/dashboards/_dashboards_prebuilds.json.tpl

@@ -111,9 +111,9 @@
           },
           "editorMode": "code",
           "expr": "min(coderd_experiments{experiment=\"workspace-prebuilds\"})",
-          "instant": false,
+          "instant": true,
           "legendFormat": "__auto",
-          "range": true,
+          "range": false,
           "refId": "A"
         }
       ],
@@ -645,7 +645,7 @@
           "refId": "E"
         }
       ],
-      "title": "Change over range: $preset",
+      "title": "Pool Capacity: $preset",
       "type": "timeseries"
     },
     {
@@ -871,7 +871,7 @@
           "refId": "F"
         }
       ],
-      "title": "Change over range: $preset",
+      "title": "Pool Operations: $preset",
       "type": "timeseries"
     },
     {

coder-observability/values.yaml

@@ -81,6 +81,11 @@ global:
             delay: 10m
             thresholds:
               notify: 1
+          UnprovisionedPrebuiltWorkspaces:
+            enabled: true
+            delay: 10m
+            thresholds:
+              warn: 1
       provisionerd:
         groups:
           Replicas: