Monitoring Octopus

Octopus is built on sigs.k8s.io/controller-runtime, so some metrics are related to controller-runtime and client-go. At the same time, github.com/prometheus/client_golang provides some metrics for Go runtime and process state.

Metrics Category

In the "Type" column, use the first letter to represent the corresponding word: G - Gauge, C - Counter, H - Histogram, S - Summary.

Exposing from Controller Runtime

Controller metrics

Type	Name	Description	Usage
C	controller_runtime_reconcile_total	Total number of reconciliations per controller.
C	controller_runtime_reconcile_errors_total	Total number of reconciliation errors per controller.
H	controller_runtime_reconcile_time_seconds	Length of time per reconciliation per controller.

Webhook metrics

Type	Name	Description	Usage
H	controller_runtime_webhook_latency_seconds	Histogram of the latency of processing admission requests.

Exposing from Kubernetes client

Rest client metrics

Type	Name	Description	Usage
C	rest_client_requests_total	Number of HTTP requests, partitioned by status code, method, and host.
H	rest_client_request_latency_seconds	Request latency in seconds. Broken down by verb and URL.

Workqueue metrics

Type	Name	Description	Usage
G	workqueue_depth	Current depth of workqueue.
G	workqueue_unfinished_work_seconds	How many seconds of work has done that is in progress and hasn't been observed by work_duration. Large values indicate stuck threads. One can deduce the number of stuck threads by observing the rate at which this increases.
G	workqueue_longest_running_processor_seconds	How many seconds has the longest running processor for workqueue been running.
C	workqueue_adds_total	Total number of adds handled by workqueue.
C	workqueue_retries_total	Total number of retries handled by workqueue.
H	workqueue_queue_duration_seconds	How long in seconds an item stays in workqueue before being requested.
H	workqueue_work_duration_seconds	How long in seconds processing an item from workqueue takes.

Exposing from Prometheus client

Go runtime metrics

Type	Name	Description	Usage
G	go_goroutines	Number of goroutines that currently exist.
G	go_threads	Number of OS threads created.
G	go_info	Information about the Go environment.
S	go_gc_duration_seconds	A summary of the pause duration of garbage collection cycles.
G	go_memstats_alloc_bytes	Number of bytes allocated and still in use.
C	go_memstats_alloc_bytes_total	Total number of bytes allocated, even if freed.
G	go_memstats_sys_bytes	Number of bytes obtained from system.
C	go_memstats_lookups_total	Total number of pointer lookups.
C	go_memstats_mallocs_total	Total number of mallocs.
C	go_memstats_frees_total	Total number of frees.
G	go_memstats_heap_alloc_bytes	Number of heap bytes allocated and still in use.
G	go_memstats_heap_sys_bytes	Number of heap bytes obtained from system.
G	go_memstats_heap_idle_bytes	Number of heap bytes waiting to be used.
G	go_memstats_heap_inuse_bytes	Number of heap bytes that are in use.
G	go_memstats_heap_released_bytes	Number of heap bytes released to OS.
G	go_memstats_heap_objects	Number of allocated objects.
G	go_memstats_stack_inuse_bytes	Number of bytes in use by the stack allocator.
G	go_memstats_stack_sys_bytes	Number of bytes obtained from system for stack allocator.
G	go_memstats_mspan_inuse_bytes	Number of bytes in use by mspan structures.
G	go_memstats_mspan_sys_bytes	Number of bytes used for mspan structures obtained from system.
G	go_memstats_mcache_inuse_bytes	Number of bytes in use by mcache structures.
G	go_memstats_mcache_sys_bytes	Number of bytes used for mcache structures obtained from system.
G	go_memstats_buck_hash_sys_bytes	Number of bytes used by the profiling bucket hash table.
G	go_memstats_gc_sys_bytes	Number of bytes used for garbage collection system metadata.
G	go_memstats_other_sys_bytes	Number of bytes used for other system allocations.
G	go_memstats_next_gc_bytes	Number of heap bytes when next garbage collection will take place.
G	go_memstats_last_gc_time_seconds	Number of seconds since 1970 of last garbage collection.
G	go_memstats_gc_cpu_fraction	The fraction of this program's available CPU time used by the GC since the program started.

Running process metrics

Type	Name	Description	Usage
C	process_cpu_seconds_total	Total user and system CPU time spent in seconds.
G	process_open_fds	Number of open file descriptors.
G	process_max_fds	Maximum number of open file descriptors.
G	process_virtual_memory_bytes	Virtual memory size in bytes.
G	process_virtual_memory_max_bytes	Maximum amount of virtual memory available in bytes.
G	process_resident_memory_bytes	Resident memory size in bytes.
G	process_start_time_seconds	Start time of the process since unix epoch in seconds.

Exposing from Octopus

Limb metrics

Type	Name	Description	Usage
G	limb_connect_connections	How many connections are connecting adaptor.
C	limb_connect_errors_total	Total number of connecting adaptor errors.
C	limb_send_errors_total	Total number of errors of sending device desired to adaptor.
H	limb_send_latency_seconds	Histogram of the latency of sending device desired to adaptor.

Monitor

By default, the metrics will be exposed on port 8080(see brain options and limb options), they can be collected by Prometheus and visually analyzed through Grafana. Octopus provides a ServiceMonitor definition YAML to integrate with Prometheus Operator, which is an easy tool to configure and manage Prometheus instances.

Grafana Dashboard

For convenience, Octopus provides a Grafana Dashboard to visualize the monitoring metrics.

;

Integrate with Prometheus Operator

Using prometheus-operator HELM chart, you can easily set up a Prometheus Operator to monitor the Octopus. The following steps demonstrate how to run a Prometheus Operator on a local Kubernetes cluster:

1. Use cluster-k3d-spinup.sh to set up a local Kubernetes cluster via k3d.
2. Follow the installation guide of HELM to install helm tool, and then use helm fetch --untar --untardir /tmp stable/prometheus-operator the prometheus-operator chart to local /tmp directory.
3. Generate a deployment YAML from prometheus-operator chart as below.
   helm template --namespace octopus-monitoring \

   --name octopus \

   --set defaultRules.create=false \

   --set global.rbac.pspEnabled=false \

   --set prometheusOperator.admissionWebhooks.patch.enabled=false \

   --set prometheusOperator.admissionWebhooks.enabled=false \

   --set prometheusOperator.kubeletService.enabled=false \

   --set prometheusOperator.tlsProxy.enabled=false \

   --set prometheusOperator.serviceMonitor.selfMonitor=false \

   --set alertmanager.enabled=false \

   --set grafana.defaultDashboardsEnabled=false \

   --set coreDns.enabled=false \

   --set kubeApiServer.enabled=false \

   --set kubeControllerManager.enabled=false \

   --set kubeEtcd.enabled=false \

   --set kubeProxy.enabled=false \

   --set kubeScheduler.enabled=false \

   --set kubeStateMetrics.enabled=false \

   --set kubelet.enabled=false \

   --set nodeExporter.enabled=false \

   --set prometheus.serviceMonitor.selfMonitor=false \

   --set prometheus.ingress.enabled=true \

   --set prometheus.ingress.hosts={localhost} \

   --set prometheus.ingress.paths={/prometheus} \

   --set prometheus.ingress.annotations.'traefik\.ingress\.kubernetes\.io\/rewrite-target'=/ \

   --set prometheus.prometheusSpec.externalUrl=http://localhost/prometheus \

   --set prometheus.prometheusSpec.serviceMonitorSelectorNilUsesHelmValues=false \

   --set prometheus.prometheusSpec.podMonitorSelectorNilUsesHelmValues=false \

   --set prometheus.prometheusSpec.ruleSelectorNilUsesHelmValues=false \

   --set grafana.adminPassword=admin \

   --set grafana.rbac.pspUseAppArmor=false \

   --set grafana.rbac.pspEnabled=false \

   --set grafana.serviceMonitor.selfMonitor=false \

   --set grafana.testFramework.enabled=false \

   --set grafana.ingress.enabled=true \

   --set grafana.ingress.hosts={localhost} \

   --set grafana.ingress.path=/grafana \

   --set grafana.ingress.annotations.'traefik\.ingress\.kubernetes\.io\/rewrite-target'=/ \

   --set grafana.'grafana\.ini'.server.root_url=http://localhost/grafana \

   /tmp/prometheus-operator > /tmp/prometheus-operator_all_in_one.yaml
4. Create octopus-monitoring Namespace via kubectl create ns octopus-monitoring.
5. Apply the prometheus-operator all-in-one deployment into the local cluster via kubectl apply -f /tmp/prometheus-operator_all_in_one.yaml.
6. Apply the Octopus all-in-one deployment via kubectl apply -f https://raw.githubusercontent.com/cnrancher/octopus/master/deploy/e2e/all_in_one.yaml.
7. Apply the monitoring integration into the local cluster via kubectl apply -f https://raw.githubusercontent.com/cnrancher/octopus/master/deploy/e2e/integrate_with_prometheus_operator.yaml
8. Visit http://localhost/prometheus to view the Prometheus web console through the browser, or visit http://localhost/grafana to view the Grafana console(the administrator account is admin/admin).
9. (Optional) Import the Octopus Overview dashboard from Grafana console.