Documentation ¶
Overview ¶
Package client contains functions to create Cadence clients used to communicate to Cadence service.
Use these to perform CRUD on domains and start or query workflow executions.
Index ¶
- Constants
- func IsWorkflowError(err error) bool
- func NewValue(data []byte) encoded.Value
- func NewValues(data []byte) encoded.Values
- type Client
- type DomainClient
- type FeatureFlags
- type HistoryEventIterator
- type Options
- type ParentClosePolicy
- type QueryWorkflowWithOptionsRequest
- type QueryWorkflowWithOptionsResponse
- type StartWorkflowOptions
- type WorkflowIDReusePolicy
- type WorkflowRun
Constants ¶
const ( // QueryTypeStackTrace is the build in query type for Client.QueryWorkflow() call. Use this query type to get the call // stack of the workflow. The result will be a string encoded in the encoded.Value. QueryTypeStackTrace string = internal.QueryTypeStackTrace // QueryTypeOpenSessions is the build in query type for Client.QueryWorkflow() call. Use this query type to get all open // sessions in the workflow. The result will be a list of SessionInfo encoded in the encoded.Value. QueryTypeOpenSessions string = internal.QueryTypeOpenSessions )
const ( // ParentClosePolicyTerminate means terminating the child workflow ParentClosePolicyTerminate = internal.ParentClosePolicyTerminate // ParentClosePolicyRequestCancel means requesting cancellation on the child workflow ParentClosePolicyRequestCancel = internal.ParentClosePolicyRequestCancel // ParentClosePolicyAbandon means not doing anything on the child workflow ParentClosePolicyAbandon = internal.ParentClosePolicyAbandon )
Variables ¶
This section is empty.
Functions ¶
func IsWorkflowError ¶ added in v0.18.4
IsWorkflowError returns true if an error is a known returned-by-the-workflow error type, when it comes from WorkflowRun from either Client.ExecuteWorkflow or Client.GetWorkflow. If it returns false, the error comes from some other source, e.g. RPC request failures or bad arguments. Using it on errors from any other source is currently undefined.
This checks for known types via errors.As, so it will check recursively if it encounters wrapped errors. Note that this is different than the various `cadence.Is[Type]Error` checks, which do not unwrap.
Currently the complete list of errors checked is: *cadence.CustomError, *cadence.CanceledError, *workflow.ContinueAsNewError, *workflow.GenericError, *workflow.TimeoutError, *workflow.TerminatedError, *workflow.PanicError, *workflow.UnknownExternalWorkflowExecutionError
See documentation for each error type for details.
func NewValue ¶ added in v0.6.1
NewValue creates a new encoded.Value which can be used to decode binary data returned by Cadence. For example: User had Activity.RecordHeartbeat(ctx, "my-heartbeat") and then got response from calling Client.DescribeWorkflowExecution. The response contains binary field PendingActivityInfo.HeartbeatDetails, which can be decoded by using:
var result string // This need to be same type as the one passed to RecordHeartbeat NewValue(data).Get(&result)
func NewValues ¶ added in v0.6.1
NewValues creates a new encoded.Values which can be used to decode binary data returned by Cadence. For example: User had Activity.RecordHeartbeat(ctx, "my-heartbeat", 123) and then got response from calling Client.DescribeWorkflowExecution. The response contains binary field PendingActivityInfo.HeartbeatDetails, which can be decoded by using:
var result1 string var result2 int // These need to be same type as those arguments passed to RecordHeartbeat NewValues(data).Get(&result1, &result2)
Types ¶
type Client ¶
type Client interface { // StartWorkflow starts a workflow execution // The user can use this to start using a function or workflow type name. // Either by // StartWorkflow(ctx, options, "workflowTypeName", arg1, arg2, arg3) // or // StartWorkflow(ctx, options, workflowExecuteFn, arg1, arg2, arg3) // The errors it can return: // - EntityNotExistsError, if domain does not exists // - BadRequestError // - WorkflowExecutionAlreadyStartedError // - InternalServiceError StartWorkflow(ctx context.Context, options StartWorkflowOptions, workflowFunc interface{}, args ...interface{}) (*workflow.Execution, error) // ExecuteWorkflow starts a workflow execution and return a WorkflowRun instance and error // The user can use this to start using a function or workflow type name. // Either by // ExecuteWorkflow(ctx, options, "workflowTypeName", arg1, arg2, arg3) // or // ExecuteWorkflow(ctx, options, workflowExecuteFn, arg1, arg2, arg3) // The errors it can return: // - EntityNotExistsError, if domain does not exists // - BadRequestError // - InternalServiceError // // WorkflowRun has 2 methods: // - GetRunID() string: which return the first started workflow run ID (please see below) // - Get(ctx context.Context, valuePtr interface{}) error: which will fill the workflow // execution result to valuePtr, if workflow execution is a success, or return corresponding // error. This is a blocking API. // NOTE: if the started workflow return ContinueAsNewError during the workflow execution, the // return result of GetRunID() will be the started workflow run ID, not the new run ID caused by ContinueAsNewError, // however, Get(ctx context.Context, valuePtr interface{}) will return result from the run which did not return ContinueAsNewError. // Say ExecuteWorkflow started a workflow, in its first run, has run ID "run ID 1", and returned ContinueAsNewError, // the second run has run ID "run ID 2" and return some result other than ContinueAsNewError: // GetRunID() will always return "run ID 1" and Get(ctx context.Context, valuePtr interface{}) will return the result of second run. // NOTE: DO NOT USE THIS API INSIDE A WORKFLOW, USE workflow.ExecuteChildWorkflow instead ExecuteWorkflow(ctx context.Context, options StartWorkflowOptions, workflow interface{}, args ...interface{}) (WorkflowRun, error) // GetWorkflow retrieves a workflow execution and return a WorkflowRun instance (described above) // - workflow ID of the workflow. // - runID can be default(empty string). if empty string then it will pick the last running execution of that workflow ID. // // WorkflowRun has 2 methods: // - GetRunID() string: which return the first started workflow run ID (please see below) // - Get(ctx context.Context, valuePtr interface{}) error: which will fill the workflow // execution result to valuePtr, if workflow execution is a success, or return corresponding // error. This is a blocking API. // If workflow not found, the Get() will return EntityNotExistsError. // NOTE: if the started workflow return ContinueAsNewError during the workflow execution, the // return result of GetRunID() will be the started workflow run ID, not the new run ID caused by ContinueAsNewError, // however, Get(ctx context.Context, valuePtr interface{}) will return result from the run which did not return ContinueAsNewError. // Say ExecuteWorkflow started a workflow, in its first run, has run ID "run ID 1", and returned ContinueAsNewError, // the second run has run ID "run ID 2" and return some result other than ContinueAsNewError: // GetRunID() will always return "run ID 1" and Get(ctx context.Context, valuePtr interface{}) will return the result of second run. GetWorkflow(ctx context.Context, workflowID string, runID string) WorkflowRun // SignalWorkflow sends a signals to a workflow in execution // - workflow ID of the workflow. // - runID can be default(empty string). if empty string then it will pick the running execution of that workflow ID. // - signalName name to identify the signal. // - arg is the data (or nil) to send with the signal, which can be read with the signal channel's Receive out-arg. // The errors it can return: // - EntityNotExistsError // - InternalServiceError // - WorkflowExecutionAlreadyCompletedError SignalWorkflow(ctx context.Context, workflowID string, runID string, signalName string, arg interface{}) error // SignalWithStartWorkflow sends a signal to a running workflow. // If the workflow is not running or not found, it starts the workflow and then sends the signal in transaction. // - workflowID, signalName, signalArg are same as SignalWorkflow's parameters // - options, workflow, workflowArgs are same as StartWorkflow's parameters // The errors it can return: // - EntityNotExistsError, if domain does not exist // - BadRequestError // - InternalServiceError SignalWithStartWorkflow(ctx context.Context, workflowID string, signalName string, signalArg interface{}, options StartWorkflowOptions, workflowFunc interface{}, workflowArgs ...interface{}) (*workflow.Execution, error) // CancelWorkflow cancels a workflow in execution // - workflow ID of the workflow. // - runID can be default(empty string). if empty string then it will pick the running execution of that workflow ID. // The errors it can return: // - EntityNotExistsError // - BadRequestError // - InternalServiceError // - WorkflowExecutionAlreadyCompletedError CancelWorkflow(ctx context.Context, workflowID string, runID string) error // TerminateWorkflow terminates a workflow execution. // workflowID is required, other parameters are optional. // - workflow ID of the workflow. // - runID can be default(empty string). if empty string then it will pick the running execution of that workflow ID. // The errors it can return: // - EntityNotExistsError // - BadRequestError // - InternalServiceError // - WorkflowExecutionAlreadyCompletedError TerminateWorkflow(ctx context.Context, workflowID string, runID string, reason string, details []byte) error // GetWorkflowHistory gets history events of a particular workflow // - workflow ID of the workflow. // - runID can be default(empty string). if empty string then it will pick the last running execution of that workflow ID. // - whether use long poll for tracking new events: when the workflow is running, there can be new events generated during iteration // of HistoryEventIterator, if isLongPoll == true, then iterator will do long poll, tracking new history event, i.e. the iteration // will not be finished until workflow is finished; if isLongPoll == false, then iterator will only return current history events. // - whether return all history events or just the last event, which contains the workflow execution end result // Example:- // To iterate all events, // iter := GetWorkflowHistory(ctx, workflowID, runID, isLongPoll, filterType) // events := []*shared.HistoryEvent{} // for iter.HasNext() { // event, err := iter.Next() // if err != nil { // return err // } // events = append(events, event) // } GetWorkflowHistory(ctx context.Context, workflowID string, runID string, isLongPoll bool, filterType s.HistoryEventFilterType) HistoryEventIterator // CompleteActivity reports activity completed. // activity Execute method can return activity.ErrResultPending to // indicate the activity is not completed when it's Execute method returns. In that case, this CompleteActivity() method // should be called when that activity is completed with the actual result and error. If err is nil, activity task // completed event will be reported; if err is CanceledError, activity task cancelled event will be reported; otherwise, // activity task failed event will be reported. // An activity implementation should use GetActivityInfo(ctx).TaskToken function to get task token to use for completion. // Example:- // To complete with a result. // CompleteActivity(token, "Done", nil) // To fail the activity with an error. // CompleteActivity(token, nil, cadence.NewCustomError("reason", details) // The activity can fail with below errors ErrorWithDetails, TimeoutError, CanceledError. CompleteActivity(ctx context.Context, taskToken []byte, result interface{}, err error) error // CompleteActivityByID reports activity completed. // Similar to CompleteActivity, but may save cadence user from keeping taskToken info. // activity Execute method can return activity.ErrResultPending to // indicate the activity is not completed when it's Execute method returns. In that case, this CompleteActivityById() method // should be called when that activity is completed with the actual result and error. If err is nil, activity task // completed event will be reported; if err is CanceledError, activity task cancelled event will be reported; otherwise, // activity task failed event will be reported. // An activity implementation should use activityID provided in ActivityOption to use for completion. // domain name, workflowID, activityID are required, runID is optional. // The errors it can return: // - ErrorWithDetails // - TimeoutError // - CanceledError CompleteActivityByID(ctx context.Context, domain, workflowID, runID, activityID string, result interface{}, err error) error // RecordActivityHeartbeat records heartbeat for an activity. // taskToken - is the value of the binary "TaskToken" field of the "ActivityInfo" struct retrieved inside the activity. // details - is the progress you want to record along with heart beat for this activity. // The errors it can return: // - EntityNotExistsError // - InternalServiceError RecordActivityHeartbeat(ctx context.Context, taskToken []byte, details ...interface{}) error // RecordActivityHeartbeatByID records heartbeat for an activity. // details - is the progress you want to record along with heart beat for this activity. // The errors it can return: // - EntityNotExistsError // - InternalServiceError RecordActivityHeartbeatByID(ctx context.Context, domain, workflowID, runID, activityID string, details ...interface{}) error // ListClosedWorkflow gets closed workflow executions based on request filters. // Retrieved workflow executions are sorted by start time in descending order. // (Retrieved workflow executions could also be sorted by closed time in descending order, // if cadence server side config EnableReadFromClosedExecutionV2 is set to true.) // Note: heavy usage of this API may cause huge persistence pressure. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError ListClosedWorkflow(ctx context.Context, request *s.ListClosedWorkflowExecutionsRequest) (*s.ListClosedWorkflowExecutionsResponse, error) // ListOpenWorkflow gets open workflow executions based on request filters. // Retrieved workflow executions are sorted by start time in descending order. // Note: heavy usage of this API may cause huge persistence pressure. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError ListOpenWorkflow(ctx context.Context, request *s.ListOpenWorkflowExecutionsRequest) (*s.ListOpenWorkflowExecutionsResponse, error) // ListWorkflow gets workflow executions based on query. This API only works with ElasticSearch, // and will return BadRequestError when using Cassandra or MySQL. The query is basically the SQL WHERE clause, // examples: // - "(WorkflowID = 'wid1' or (WorkflowType = 'type2' and WorkflowID = 'wid2'))". // - "CloseTime between '2019-08-27T15:04:05+00:00' and '2019-08-28T15:04:05+00:00'". // - to list only open workflow use "CloseTime = missing" // Retrieved workflow executions are sorted by StartTime in descending order when list open workflow, // and sorted by CloseTime in descending order for other queries. // The errors it can return: // - BadRequestError // - InternalServiceError ListWorkflow(ctx context.Context, request *s.ListWorkflowExecutionsRequest) (*s.ListWorkflowExecutionsResponse, error) // ListArchivedWorkflow gets archived workflow executions based on query. This API will return BadRequest if Cadence // cluster or target domain is not configured for visibility archival or read is not enabled. The query is basically the SQL WHERE clause. // However, different visibility archivers have different limitations on the query. Please check the documentation of the visibility archiver used // by your domain to see what kind of queries are accept and whether retrieved workflow executions are ordered or not. // The errors it can return: // - BadRequestError // - InternalServiceError ListArchivedWorkflow(ctx context.Context, request *s.ListArchivedWorkflowExecutionsRequest) (*s.ListArchivedWorkflowExecutionsResponse, error) // ScanWorkflow gets workflow executions based on query. This API only works with ElasticSearch, // and will return BadRequestError when using Cassandra or MySQL. The query is basically the SQL WHERE clause // (see ListWorkflow for query examples). // ScanWorkflow should be used when retrieving large amount of workflows and order is not needed. // It will use more ElasticSearch resources than ListWorkflow, but will be several times faster // when retrieving millions of workflows. // The errors it can return: // - BadRequestError // - InternalServiceError ScanWorkflow(ctx context.Context, request *s.ListWorkflowExecutionsRequest) (*s.ListWorkflowExecutionsResponse, error) // CountWorkflow gets number of workflow executions based on query. This API only works with ElasticSearch, // and will return BadRequestError when using Cassandra or MySQL. The query is basically the SQL WHERE clause // (see ListWorkflow for query examples). // The errors it can return: // - BadRequestError // - InternalServiceError CountWorkflow(ctx context.Context, request *s.CountWorkflowExecutionsRequest) (*s.CountWorkflowExecutionsResponse, error) // GetSearchAttributes returns valid search attributes keys and value types. // The search attributes can be used in query of List/Scan/Count APIs. Adding new search attributes requires cadence server // to update dynamic config ValidSearchAttributes. GetSearchAttributes(ctx context.Context) (*s.GetSearchAttributesResponse, error) // QueryWorkflow queries a given workflow's last execution and returns the query result synchronously. Parameter workflowID // and queryType are required, other parameters are optional. The workflowID and runID (optional) identify the // target workflow execution that this query will be send to. If runID is not specified (empty string), server will // use the currently running execution of that workflowID. The queryType specifies the type of query you want to // run. By default, cadence supports "__stack_trace" as a standard query type, which will return string value // representing the call stack of the target workflow. The target workflow could also setup different query handler // to handle custom query types. // See comments at workflow.SetQueryHandler(ctx Context, queryType string, handler interface{}) for more details // on how to setup query handler within the target workflow. // - workflowID is required. // - runID can be default(empty string). if empty string then it will pick the running execution of that workflow ID. // - queryType is the type of the query. // - args... are the optional query parameters, which will be sent to your query handler function's args. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError // - QueryFailError QueryWorkflow(ctx context.Context, workflowID string, runID string, queryType string, args ...interface{}) (encoded.Value, error) // QueryWorkflowWithOptions queries a given workflow execution and returns the query result synchronously. // See QueryWorkflowWithOptionsRequest and QueryWorkflowWithOptionsResponse for more information. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError // - QueryFailError QueryWorkflowWithOptions(ctx context.Context, request *QueryWorkflowWithOptionsRequest) (*QueryWorkflowWithOptionsResponse, error) // ResetWorkflow reset a given workflow execution and returns a new execution // See ResetWorkflowRequest and ResetWorkflowResponse for more information. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError ResetWorkflow(ctx context.Context, request *s.ResetWorkflowExecutionRequest) (*s.ResetWorkflowExecutionResponse, error) // DescribeWorkflowExecution returns information about the specified workflow execution. // - runID can be default(empty string). if empty string then it will pick the last running execution of that workflow ID. // // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError DescribeWorkflowExecution(ctx context.Context, workflowID, runID string) (*s.DescribeWorkflowExecutionResponse, error) // DescribeTaskList returns information about the target tasklist, right now this API returns the // pollers which polled this tasklist in last few minutes. // The errors it can return: // - BadRequestError // - InternalServiceError // - EntityNotExistError DescribeTaskList(ctx context.Context, tasklist string, tasklistType s.TaskListType) (*s.DescribeTaskListResponse, error) }
Client is the client for starting and getting information about a workflow executions as well as completing activities asynchronously.
type DomainClient ¶
type DomainClient interface { // Register a domain with cadence server // The errors it can throw: // - DomainAlreadyExistsError // - BadRequestError // - InternalServiceError Register(ctx context.Context, request *s.RegisterDomainRequest) error // Describe a domain. The domain has 3 part of information // DomainInfo - Which has Name, Status, Description, Owner Email // DomainConfiguration - Configuration like Workflow Execution Retention Period In Days, Whether to emit metrics. // ReplicationConfiguration - replication config like clusters and active cluster name // The errors it can throw: // - EntityNotExistsError // - BadRequestError // - InternalServiceError Describe(ctx context.Context, name string) (*s.DescribeDomainResponse, error) // Update a domain. // The errors it can throw: // - EntityNotExistsError // - BadRequestError // - InternalServiceError Update(ctx context.Context, request *s.UpdateDomainRequest) error }
DomainClient is the client for managing operations on the domain. CLI, tools, ... can use this layer to manager operations on domain.
func NewDomainClient ¶
func NewDomainClient(service workflowserviceclient.Interface, options *Options) DomainClient
NewDomainClient creates an instance of a domain client, to manage lifecycle of domains.
type FeatureFlags ¶ added in v0.18.1
type FeatureFlags = internal.FeatureFlags
FeatureFlags define which breaking changes can be enabled for client
type HistoryEventIterator ¶
type HistoryEventIterator = internal.HistoryEventIterator
HistoryEventIterator is a iterator which can return history events
type Options ¶
type Options = internal.ClientOptions
Options are optional parameters for Client creation.
type ParentClosePolicy ¶ added in v0.10.0
type ParentClosePolicy = internal.ParentClosePolicy
ParentClosePolicy defines the behavior performed on a child workflow when its parent is closed
type QueryWorkflowWithOptionsRequest ¶ added in v0.9.3
type QueryWorkflowWithOptionsRequest = internal.QueryWorkflowWithOptionsRequest
QueryWorkflowWithOptionsRequest defines the request to QueryWorkflowWithOptions
type QueryWorkflowWithOptionsResponse ¶ added in v0.9.3
type QueryWorkflowWithOptionsResponse = internal.QueryWorkflowWithOptionsResponse
QueryWorkflowWithOptionsResponse defines the response to QueryWorkflowWithOptions
type StartWorkflowOptions ¶
type StartWorkflowOptions = internal.StartWorkflowOptions
StartWorkflowOptions configuration parameters for starting a workflow execution.
type WorkflowIDReusePolicy ¶
type WorkflowIDReusePolicy = internal.WorkflowIDReusePolicy
WorkflowIDReusePolicy defines workflow ID reuse behavior.
const ( // WorkflowIDReusePolicyAllowDuplicateFailedOnly allow start a workflow execution // when workflow not running, and the last execution close state is in // [terminated, cancelled, timeout, failed]. WorkflowIDReusePolicyAllowDuplicateFailedOnly WorkflowIDReusePolicy = internal.WorkflowIDReusePolicyAllowDuplicateFailedOnly // WorkflowIDReusePolicyAllowDuplicate allow start a workflow execution using // the same workflow ID,when workflow not running. WorkflowIDReusePolicyAllowDuplicate WorkflowIDReusePolicy = internal.WorkflowIDReusePolicyAllowDuplicate // WorkflowIDReusePolicyRejectDuplicate do not allow start a workflow execution using the same workflow ID at all WorkflowIDReusePolicyRejectDuplicate WorkflowIDReusePolicy = internal.WorkflowIDReusePolicyRejectDuplicate // WorkflowIDReusePolicyTerminateIfRunning terminate current running workflow using the same workflow ID if exist, // then start a new run in one transaction WorkflowIDReusePolicyTerminateIfRunning = internal.WorkflowIDReusePolicyTerminateIfRunning )
type WorkflowRun ¶
type WorkflowRun = internal.WorkflowRun
WorkflowRun represents a started non child workflow