ActionType refers to a category of actions that can be associated with a specific controller.

Name is a unique identifier of the action

DependsOn defines the conditions for the execution of this action

Assert defines the conditions under which the Scenario will terminate with a “passed” or “failed” message

(Members of EmbedActions are embedded into this type.)

"Call"

ActionCall starts a remote process execution, from the controller to the targeted services.

"Cascade"

ActionCascade injects multiple failures into the running system.

"Chaos"

ActionChaos injects failures into the running system.

"Cluster"

ActionCluster creates multiple services running in a shared context.

"Delete"

ActionDelete deletes a created Frisbee resource (i.e services, clusters,).

"Service"

ActionService creates a new service.

Callable is the name of the endpoint that will be called

Services is a list of services that will be stopped.

Until defines the conditions under which the CR will stop spawning new jobs. If used in conjunction with inputs, it will loop over inputs until the conditions are met.

Schedule defines the interval between the invocations of the callable.

Suspend flag tells the controller to suspend subsequent executions, it does not apply to already started executions. Defaults to false.

(Members of Lifecycle are embedded into this type.)

QueuedJobs is a list of services scheduled for stopping.

ScheduledJobs points to the next QueuedJobs.

LastScheduleTime provide information about the last time a Service was successfully scheduled.

Container specific the name of the container to which we will run the command

Container specifies a command and arguments to stop the targeted container in an application-specific manner.

(Members of GenerateFromTemplate are embedded into this type.)

Schedule defines the interval between the creation of services within the group.

Suspend flag tells the controller to suspend subsequent executions, it does not apply to already started executions. Defaults to false.

(Members of Lifecycle are embedded into this type.)

QueuedJobs is a list of Chaos jobs scheduled for creation by the cascade.

ScheduledJobs points to the next QueuedJobs.

LastScheduleTime provide information about the last time a Chaos job was successfully scheduled.

Type indicate the type of the injected fault

(Members of EmbedFaultType are embedded into this type.)

(Members of Lifecycle are embedded into this type.)

LastScheduleTime provide information about the last time a Pod was scheduled.

(Members of GenerateFromTemplate are embedded into this type.)

Tolerate specifies the conditions under which the cluster will fail. If undefined, the cluster fails immediately when a service has failed.

Schedule defines the interval between the creation of services in the group.

Placement defines rules for placing the containers across the available nodes.

Suspend flag tells the controller to suspend subsequent executions, it does not apply to already started executions. Defaults to false.

(Members of Lifecycle are embedded into this type.)

QueuedJobs is a list of services scheduled for creation by the cluster.

ScheduledJobs points to the next QueuedJobs.

LastScheduleTime provide information about the last time a Service was successfully scheduled.

"AllJobsAreCompleted"

ConditionAllJobsAreCompleted indicates that all jobs have been successfully completed. Jobs may refer to actions of a scenario, services of a cluster, chaos events of a cascade, etc.

"AllJobsAreScheduled"

ConditionAllJobsAreScheduled indicates that all jobs have been successfully scheduled. Jobs may refer to actions of a scenario, services of a cluster, chaos events of a cascade, etc.

"Initialized"

ConditionCRInitialized indicates whether the workflow has been initialized

"UnexpectedTermination"

ConditionJobUnexpectedTermination is used for a job that has been unexpectedly terminated. The termination refers to both Success and Fail.

"Terminated"

ConditionTerminated indicates the user-defined conditions are met.

Metrics set a Grafana alert that will be triggered once the condition is met. Parsing: Grafana URL: http://grafana/d/A2EjFbsMk/ycsb-services?editPanel=86 metrics: A2EjFbsMk/86/Average (Panel/Dashboard/Metric)

State describe the runtime condition that should be met after the action has been executed Shall be defined using .Lifecycle() methods. The methods account only jobs that are managed by the object.

SetFields is used to populate fields. Used for dynamic assignment based templated inputs.

Resources specifies limitations as to how the container will access host resources.

Telemetry is a list of referenced agents responsible to monitor the Service. Agents are sidecar services will be deployed in the same Pod as the Service container.

Jobs is a list of jobs to be deleted. The format is {“kind”:“name”}, e.g, {“service”,“client”}

ReadBPS limits read rate (bytes per second)

ReadIOPS limits read rate (IO per second)

WriteBPS limits write rate (bytes per second)

WriteIOPS limits write rate (IO per second)

TemplateRef refers to a template (e.g, iperf-server).

Until defines the conditions under which the CR will stop spawning new jobs. If used in conjunction with inputs, it will loop over inputs until the conditions are met.

MaxInstances dictate the number of objects to be created for the CR. If no inputs are defined, then all instances will be initiated using the default parameters of the template. Conditions used in conjunction with Until, MaxInstances as a max bound.

Inputs are list of inputs passed to the objects. Conditions used in conjunction with instances, if the number of instances is larger that the number of inputs, then inputs are recursively iteration.

Parameters are user-set values that are dynamically evaluated

Phase is a simple, high-level summary of where the Object is in its lifecycle. The conditions array, the reason and message fields, and the individual container status arrays contain more detail about the pod’s status.

Reason is A brief CamelCase message indicating details about why the service is in this Phase. e.g. ‘Evicted’

Message provides more details for understanding the Reason.

Conditions describe sequences of events that warrant the present Phase.

ByName is a map of string keys and a set values that used to select services. The key defines the namespace which services belong, and the values is a set of service names.

ByCluster defines the service group where services belong.

"all"

AllMode represents that the system will select all objects regardless of status (not ready or not running pods includes). Use this label carefully.

"fixed"

FixedMode represents that the system will select a specific number of running objects.

"fixed-percent"

FixedPercentMode to specify a fixed % of a cluster

"one"

OneMode represents that the system will select one object randomly.

"random-max-percent"

RandomMaxPercentMode to specify a maximum % of a cluster

"Failed"

PhaseFailed means that at least one job of the CR has terminated in a failure (exited with a non-zero exit code or was stopped by the system).

"Pending"

PhasePending means that the CR has been accepted by the Kubernetes cluster, but one of the child jobs has not been created. This includes the time waiting for logical dependencies, Ports discovery, data rewiring, and placement of Pods.

"Running"

PhaseRunning means that all the child jobs of a CR have been created, and at least one job is still running.

"Success"

PhaseSuccess means that all jobs in a CR have voluntarily exited, and the system is not going to restart any of these Jobs.

""

PhaseUninitialized means that request is not yet accepted by the controller.

Collocate will place all the services within the same node.

ConflictsWith points to another cluster whose services cannot be located with this one. For example, this is needed for placing the master nodes on a different failure domain than the slave nodes.

Nodes will place all the services within the same specific node.

EphemeralVolume creates an ephemeral volume type.

Ingress makes An API object that manages external access to the services in a cluster, typically HTTP.

Cron defines a cron job rule.

Some rule examples: “0 30 * * * *” means to “Every hour on the half hour” “@hourly” means to “Every hour” “@every 1h30m” means to “Every hour thirty”

More rule info: https://godoc.org/github.com/robfig/cron

StartingDeadlineSeconds is an optional deadline in seconds for starting the job if it misses scheduled time for any reason. if we miss this deadline, we’ll just wait till the next scheduled time

Match contains the rules to select target

Mode defines which of the selected services to use. If undefined, all() is used Supported mode: one / all / fixed / fixed-percent / random-max-percent

Value is required when the mode is set to FixedPodMode / FixedPercentPodMod / RandomMaxPercentPodMod. If FixedPodMode, provide an integer of pods to do chaos action. If FixedPercentPodMod, provide a number from 0-100 to specify the percent of pods the server can do chaos action. IF RandomMaxPercentPodMod, provide a number from 0-100 to specify the max percent of pods to do chaos action

Macro abstract selector parameters into a structured string (e.g, .cluster.master.all). Every parsed field is represents an inner structure of the selector. In case of invalid macro, the selector will return empty results. Macro conflicts with any other parameter.

(Members of PodSpec are embedded into this type.)

(Members of Lifecycle are embedded into this type.)

LastScheduleTime provide information about the last time a Pod was scheduled.

Field is the path to the field whose value will be replaced. Examples: Containers.0.Ports.0

Inputs are dynamic fields that populate the spec.

(Members of EmbedSpecs are embedded into this type.)

EmbedSpecs point to the Frisbee specs that can be templated.

(Members of Lifecycle are embedded into this type.)

Actions are the tasks that will be taken.

Suspend flag tells the controller to suspend subsequent executions, it does not apply to already started executions. Defaults to false.

(Members of Lifecycle are embedded into this type.)

ExecutedActions is a list of executed actions.

GrafanaEndpoint points to the local Grafana instance

PrometheusEndpoint points to the local Prometheus instance

FailedServices indicate the number of services that may fail before the cluster fails itself.

(Members of Lifecycle are embedded into this type.)

Running waits for the given groups to be running

Success waits for the given groups to be succeeded

After is the time offset since the beginning of this action.