feat: linting and end-to-end collection tests (#28)

.github/workflows/test.yml

@@ -0,0 +1,51 @@
+name: Test Charts
+
+on: pull_request
+
+jobs:
+  lint-charts:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Install Helm
+        uses: azure/setup-helm@v3
+        with:
+          version: v3.12.1
+
+      - uses: actions/setup-python@v4
+        with:
+          python-version: '3.9'
+          check-latest: true
+
+      - name: Install chart-testing
+        uses: helm/[email protected]
+        with:
+          version: v3.8.0
+
+      - name: Lint charts
+        run: make lint
+
+  test-charts:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Install Helm
+        uses: azure/setup-helm@v3
+        with:
+          version: v3.12.1
+
+      - name: Install kind
+        uses: helm/[email protected]
+        with:
+          install_only: true
+
+      - name: Test charts
+        run: make test GOBUILD=false

.gitignore

@@ -1 +1,3 @@
 *.tgz
+demo
+.kubeconfig

Makefile

@@ -1,17 +1,49 @@
-CHARTS := $(shell dirname `find . -name Chart.yaml`)
+# Charts will have their dependencies built in the order defined here. Endpoint is
+# a dependency of all charts, and should thus always be first in the list. Stack
+# and traces are umbrella charts and should always be at the end of the list.
+CHARTS := endpoint proxy events logs metrics stack traces
+
+.PHONY: all
+all: build-deps lint test
+
+add-repos:
+	@helm repo add observe https://observeinc.github.io/helm-charts
+	@helm repo add grafana https://grafana.github.io/helm-charts
+	@helm repo add fluent https://fluent.github.io/helm-charts
+	@helm repo add otel https://open-telemetry.github.io/opentelemetry-helm-charts
+	helm repo up
 
 .PHONY: update-deps
-update-deps:
+update-deps: add-repos
 	@for chart in $(CHARTS); do \
-		helm dependency update $$chart; \
+		helm dependency update --skip-refresh charts/$$chart; \
 	done
 
 .PHONY: build-deps
-build-deps:
-	helm repo add grafana https://grafana.github.io/helm-charts
-	helm repo add fluent https://fluent.github.io/helm-charts
-	helm repo add otel https://open-telemetry.github.io/opentelemetry-helm-charts
-	helm repo update
+build-deps: add-repos
+	@for chart in $(CHARTS); do \
+		echo building chart dependencies for charts/$$chart...; \
+		helm dependency build --skip-refresh charts/$$chart; \
+		echo ; \
+	done
+
+.PHONY: test
+test: build-deps build-test-images
+	test/test.sh stack traces
+
+.PHONY: lint
+lint: build-deps
+	ct lint --all --helm-dependency-extra-args='--skip-refresh'
+
+.PHONY: clean
+clean:
+	test/clean.sh
 	@for chart in $(CHARTS); do \
-		helm dependency build --skip-refresh $$chart; \
+		echo rm -f charts/$$chart/charts/*.tgz; \
+		rm -f charts/$$chart/charts/*.tgz; \
 	done
+	make -C test/cmd clean
+
+.PHONY: build-test-images
+build-test-images:
+	make -C test/cmd all

charts/internal/endpoint/Chart.yaml → charts/endpoint/Chart.yaml

@@ -3,3 +3,6 @@ name: endpoint
 description: Observe collection endpoint utility functions
 type: library
 version: 0.1.3
+maintainers:
+  - name: Observe
+    email: [email protected]

charts/internal/endpoint/templates/_address.tpl → charts/endpoint/templates/_address.tpl

@@ -1,6 +1,8 @@
 
-# If a collection endpoint is provided, parse the scheme from that.
-# Otherwise, fall back to legacy collectorScheme value.
+{{- /*
+If a collection endpoint is provided, parse the scheme from that.
+Otherwise, fall back to legacy collectorScheme value.
+*/}}
 {{- define "observe.collectorScheme" -}}{{- with .Values.global.observe -}}
     {{- if .collectionEndpoint -}}
         {{- with urlParse .collectionEndpoint -}}
@@ -11,9 +13,11 @@
     {{- end -}}
 {{- end -}}{{- end -}}
 
-# If a collection endpoint is provided, parse the host from that and split the host from the port.
-# Otherwise, use the legacy collectorHost value, the the customer part of the hostname prepended:
-# <customer_id>.<collectorHost>
+{{- /*
+If a collection endpoint is provided, parse the host from that and split the host from the port.
+Otherwise, use the legacy collectorHost value, the the customer part of the hostname prepended:
+<customer_id>.<collectorHost>
+*/}}
 {{- define "observe.collectorHost" -}}{{- with .Values.global.observe -}}
     {{- if .collectionEndpoint -}}
         {{- with urlParse .collectionEndpoint -}}
@@ -25,9 +29,11 @@
     {{- end -}}
 {{- end -}}{{- end -}}
 
-# If a collection endpoint is provided, look for a provided port and otherwise try to guess
-# based on the scheme.
-# If an endpoint is not provided, fall back to legacy collectorPort.
+{{- /*
+If a collection endpoint is provided, look for a provided port and otherwise try to guess
+based on the scheme.
+If an endpoint is not provided, fall back to legacy collectorPort.
+*/}}
 {{- define "observe.collectorPort" -}}{{- with .Values.global.observe -}}
     {{- if .collectionEndpoint -}}
         {{- with urlParse .collectionEndpoint -}}
@@ -47,8 +53,10 @@
     {{- end -}}
 {{- end -}}{{- end -}}
 
-# Return true if the scheme is "https", as parsed from the collection endpoint.
-# If a collection endpoint is not provided, look at the legacy collectorScheme value.
+{{- /*
+Return true if the scheme is "https", as parsed from the collection endpoint.
+If a collection endpoint is not provided, look at the legacy collectorScheme value.
+*/}}
 {{- define "observe.useTLS" -}}{{- with .Values.global.observe -}}
     {{- if .collectionEndpoint -}}
        {{- if eq "http" (urlParse .collectionEndpoint).scheme -}}
@@ -65,9 +73,11 @@
     {{- end -}}
 {{- end -}}{{- end -}}
 
-# If a collection endpoint is provided, use that.
-# Otherwise, generate the endpoint using the legacy values.
-# Re-constructing the parsed URL eliminates any path that was included in the value.
+{{- /*
+If a collection endpoint is provided, use that.
+Otherwise, generate the endpoint using the legacy values.
+Re-constructing the parsed URL eliminates any path that was included in the value.
+*/}}
 {{- define "observe.collectionEndpoint" -}}
     {{- if .Values.global.observe.collectionEndpoint -}}
         {{- with urlParse .Values.global.observe.collectionEndpoint -}}
@@ -78,7 +88,9 @@
     {{- end -}}
 {{- end -}}
 
-# Same as "observe.collectionEndpoint", but with the token provided as part of the URL.
+{{- /*
+mychart.shortname provides a 6 char truncated version of the release name.
+*/}}
 {{- define "observe.collectionEndpointWithToken" -}}
     {{- if .Values.global.observe.collectionEndpoint -}}
         {{- with urlParse .Values.global.observe.collectionEndpoint -}}

charts/endpoint/values.yaml

@@ -0,0 +1,9 @@
+global:
+  observe:
+    collectionEndpoint: ""
+
+    # Legacy configuration values, only used if collectionEndpoint is not set
+    customer: ""
+    collectorScheme: "https"
+    collectorHost: "collect.observeinc.com"
+    collectorPort: "443"

charts/internal/events/Chart.yaml → charts/events/Chart.yaml

@@ -8,3 +8,6 @@ dependencies:
   - name: endpoint
     version: 0.1.3
     repository: file://../endpoint
+maintainers:
+  - name: Observe
+    email: [email protected]

charts/internal/events/values.yaml → charts/events/values.yaml

@@ -1,3 +1,7 @@
+global:
+  # https://github.com/helm/helm/issues/8489 workaround
+  observe: {}
+
 image:
   kube_cluster_info:
     repository: observeinc/kube-cluster-info

charts/internal/logs/Chart.lock → charts/logs/Chart.lock

@@ -6,4 +6,4 @@ dependencies:
   repository: file://../endpoint
   version: 0.1.3
 digest: sha256:53c32af7f65d7683aed0aa325738388ab0be6379c25b6a0d90724e520050f54b
-generated: "2023-07-11T11:35:39.457678-07:00"
+generated: "2023-07-20T15:01:44.474585-07:00"

charts/internal/logs/Chart.yaml → charts/logs/Chart.yaml

@@ -10,3 +10,6 @@ dependencies:
   - name: endpoint
     version: 0.1.3
     repository: file://../endpoint
+maintainers:
+  - name: Observe
+    email: [email protected]

charts/internal/logs/values.yaml → charts/logs/values.yaml

@@ -1,3 +1,7 @@
+global:
+  # https://github.com/helm/helm/issues/8489 workaround
+  observe: {}
+
 fluent-bit:
   nameOverride: logs
   env:
@@ -39,8 +43,6 @@ fluent-bit:
     buffer_max_size: 256k
     dns_mode: UDP
     dns_resolver: LEGACY
-    docker_mode: On
-    docker_mode_flush: 4
     flush: 2
     grace: 10
     hc_errors_count: 5
@@ -50,14 +52,14 @@ fluent-bit:
     mem_buf_limit: 10MB
     node_log_include_path: /var/log/kube-apiserver-audit.log
     node_log_exclude_path: nomatch
-    read_from_head: True
+    read_from_head: "true"
     refresh_interval: 2
     retry_limit: 5
     rotate_wait: 5
-    storage_metrics: Off
+    storage_metrics: "off"
     grep_match_tag: "nothing"
     grep_exclude: "nomatch ^$"
-    inotify_watcher: true
+    inotify_watcher: "true"
 
     service: |
       [SERVICE]
@@ -170,7 +172,7 @@ fluent-bit:
           Header              Authorization Bearer ${OBSERVE_TOKEN}
           Compress            gzip
           Retry_Limit         {{.Values.config.retry_limit}}
-    
+
     customParsers: |
       [PARSER]
           Name        kube-custom

charts/internal/metrics/Chart.yaml → charts/metrics/Chart.yaml

@@ -10,3 +10,6 @@ dependencies:
   - name: endpoint
     version: 0.1.3
     repository: file://../endpoint
+maintainers:
+  - name: Observe
+    email: [email protected]

charts/internal/metrics/values.yaml → charts/metrics/values.yaml

@@ -1,44 +1,48 @@
+global:
+  # https://github.com/helm/helm/issues/8489 workaround
+  observe: {}
+
 grafana-agent:
   nameOverride: metrics
   prom_config:
-      batch_send_deadline: 5s
-      capacity: 15000
-      log_level: info
-      max_backoff: 30s
-      max_samples_per_send: 5000
-      max_shards: 10
-      max_wal_time: 30m
-      min_backoff: 1s
-      min_wal_time: 15s
-      observe_collector_insecure: false
-      remote_timeout: 30s
-      remote_flush_deadline: 1m
-      scrape_interval: 15s
-      scrape_timeout: 10s
-      scrape_body_size_limit: 50MB
-      scrape_cadvisor_action: keep
-      host_filter: "false"
-      # the following metrics are exported with 0 values in default cadvisor installs
-      # see
-      # - https://github.com/kubernetes/kubernetes/issues/60279
-      # - https://github.com/google/cadvisor/issues/1672
-      scrape_cadvisor_metric_drop_regex: container_(network_tcp_usage_total|network_udp_usage_total|tasks_state|cpu_load_average_10s)
-      # these are the metrics we use for our default boards
-      scrape_cadvisor_metric_keep_regex: container_(cpu_cfs_.*|spec_.*|cpu_cores|cpu_usage_seconds_total|memory_working_set_bytes|memory_usage_bytes|network_transmit_.*|network_receive_.*|fs_writes_total|fs_reads_total|file_descriptors)|machine_(cpu_cores|memory_bytes)
-      scrape_kubelet_action: drop
-      scrape_kubelet_metric_drop_regex: 
-      scrape_kubelet_metric_keep_regex: (.*)
-      scrape_pod_action: keep
-      scrape_pod_namespace_drop_regex: (.*istio.*|.*ingress.*|kube-system)
-      scrape_pod_namespace_keep_regex: (.*)
-      scrape_pod_port_keep_regex: .*metrics
-      scrape_pod_metric_drop_regex: .*bucket
-      scrape_pod_metric_keep_regex: (.*)
-      scrape_resource_action: keep
-      scrape_resource_metric_drop_regex: 
-      scrape_resource_metric_keep_regex: (.*)
-      scrape_sample_limit: 100000
-      wal_truncate_frequency: 30m
+    batch_send_deadline: 5s
+    capacity: 15000
+    log_level: info
+    max_backoff: 30s
+    max_samples_per_send: 5000
+    max_shards: 10
+    max_wal_time: 30m
+    min_backoff: 1s
+    min_wal_time: 15s
+    observe_collector_insecure: false
+    remote_timeout: 30s
+    remote_flush_deadline: 1m
+    scrape_interval: 15s
+    scrape_timeout: 10s
+    scrape_body_size_limit: 50MB
+    scrape_cadvisor_action: keep
+    host_filter: "false"
+    # the following metrics are exported with 0 values in default cadvisor installs
+    # see
+    # - https://github.com/kubernetes/kubernetes/issues/60279
+    # - https://github.com/google/cadvisor/issues/1672
+    scrape_cadvisor_metric_drop_regex: container_(network_tcp_usage_total|network_udp_usage_total|tasks_state|cpu_load_average_10s)
+    # these are the metrics we use for our default boards
+    scrape_cadvisor_metric_keep_regex: container_(cpu_cfs_.*|spec_.*|cpu_cores|cpu_usage_seconds_total|memory_working_set_bytes|memory_usage_bytes|network_transmit_.*|network_receive_.*|fs_writes_total|fs_reads_total|file_descriptors)|machine_(cpu_cores|memory_bytes)
+    scrape_kubelet_action: drop
+    scrape_kubelet_metric_drop_regex:
+    scrape_kubelet_metric_keep_regex: (.*)
+    scrape_pod_action: keep
+    scrape_pod_namespace_drop_regex: (.*istio.*|.*ingress.*|kube-system)
+    scrape_pod_namespace_keep_regex: (.*)
+    scrape_pod_port_keep_regex: .*metrics
+    scrape_pod_metric_drop_regex: .*bucket
+    scrape_pod_metric_keep_regex: (.*)
+    scrape_resource_action: keep
+    scrape_resource_metric_drop_regex:
+    scrape_resource_metric_keep_regex: (.*)
+    scrape_sample_limit: 100000
+    wal_truncate_frequency: 30m
 
   controller:
     type: deployment

charts/proxy/Chart.yaml

@@ -0,0 +1,8 @@
+apiVersion: v2
+name: proxy
+description: "Observe agent proxy (for testing only, do not use)"
+type: application
+version: 0.1.0
+maintainers:
+  - name: Observe
+    email: [email protected]