glide-mq

Classes

Class	Description
BatchError	Thrown by batch processors to report per-job results (mixed success/failure).
Broadcast	Broadcast - Fan-out message publisher for pub/sub patterns.
BroadcastWorker	-
ConnectionError	Thrown when a Valkey/Redis connection cannot be established.
CycleError	-
DelayedError	Internal control-flow error thrown by job.moveToDelayed(). Caught by the Worker.
FlowProducer	Creates parent-child job flows and DAG workflows. Children are processed first; the parent runs after all children complete.
GlideMQError	Base error class for all glide-mq errors.
GroupRateLimitError	Internal control-flow error thrown by job.rateLimitGroup(). Caught by the Worker.
Job	Represents a single job in the queue. Provides methods for reporting usage, streaming chunks, managing state, and interacting with the job lifecycle (delay, suspend, retry, etc.).
Producer	-
Queue	Queue manages job submission, retrieval, scheduling, and lifecycle operations. Connects to Valkey/Redis and uses server-side functions for atomic operations.
QueueEvents	-
ServerlessPool	-
SuspendError	Internal control-flow error thrown by job.suspend(). Caught by the Worker.
UnrecoverableError	Thrown to signal a job failure that should never be retried, regardless of attempts config.
WaitingChildrenError	Internal control-flow error thrown by job.moveToWaitingChildren(). Caught by the Worker.
Worker	-

Interfaces

Interface	Description
AddAndWaitOptions	Options for Queue.addAndWait(). Extends JobOptions with a wait timeout.
BatchOptions	Configuration for batch processing mode.
BroadcastOptions	Configuration options for a Broadcast (pub/sub) queue.
BroadcastWorkerOptions	Configuration options for a BroadcastWorker (subscriber).
BudgetOptions	Budget constraints for a flow. Caps token usage and/or cost across all jobs.
ConnectionOptions	Options for connecting to a Valkey/Redis server or cluster.
DAGFlow	A complete DAG flow definition for submission via FlowProducer.addDAG().
DAGNode	A node in a DAG flow. Each node is a job with optional dependencies on other nodes. The deps array lists the names of nodes that must complete before this node can run.
DeadLetterQueueOptions	Configuration for dead letter queue routing.
FlowJob	Definition of a job within a flow (parent-child hierarchy).
FlowProducerOptions	Configuration options for creating a FlowProducer instance.
GetJobsOptions	Options for getJob() and getJobs() methods.
GroupRateLimitOptions	Options controlling behavior when job.rateLimitGroup() is called.
IamCredentials	IAM authentication credentials for AWS ElastiCache/MemoryDB.
IndexCreateOptions	Options passed through to FT.CREATE (glide-mq owned, decoupled from speedkey).
JobCounts	Count of jobs in each state.
JobData	Generic job data payload type.
JobIndexOptions	Options for creating a Valkey Search index over job hashes.
JobNode	-
JobOptions	Options for controlling individual job behavior (delay, priority, retry, etc.).
JobTemplate	Template for jobs created by a scheduler.
JobUsage	AI-specific usage metadata reported by a job processor.
Metrics	Aggregated metrics result with total count and per-minute data points.
MetricsDataPoint	A single per-minute metrics data point.
MetricsOptions	Options for querying metrics data points.
PasswordCredentials	Standard password-based credentials.
ProducerOptions	-
QueueEventsOptions	Configuration options for creating a QueueEvents listener.
QueueOptions	Configuration options for creating a Queue instance.
RateLimitConfig	Configuration for time-window rate limiting.
ReadStreamOptions	Options for readStream() - reading entries from a job's streaming channel.
SandboxOptions	Options for sandboxed (file-path) processors running in worker threads or child processes. When a job's abort signal fires (timeout or revocation), the abort is forwarded to the processor. If it has not settled 5 seconds later, its worker thread is terminated (or its child process is SIGKILLed) and replaced, so a hung processor cannot hold a sandbox slot. During that window the worker no longer owns the job: job.updateProgress(), job.updateData() and job.moveToDelayed() reject with Job aborted; job.log() still works.
ScheduleOpts	Options for defining a job schedule (cron, interval, or repeat-after-complete).
SchedulerEntry	Stored state of a registered job scheduler.
SearchJobsOptions	Options for searchJobs() method.
SearchQueryOptions	Options passed through to FT.SEARCH (glide-mq owned, decoupled from speedkey).
Serializer	Custom serializer for job data and return values.
SignalEntry	A signal entry delivered to a suspended job on resume.
SuspendOptions	Options for suspending a job.
TokenBucketConfig	Configuration for token bucket rate limiting with burst capacity and refill rate.
UsageQueueSummary	Aggregated usage totals for a single queue inside a usage summary response.
UsageSummary	Rolling aggregate usage summary across one or more queues.
UsageSummaryOptions	Options for aggregating reported usage across one or more queues.
VectorSearchOptions	Options for vector similarity search over indexed jobs.
VectorSearchResult	A single result from a vector similarity search.
WorkerInfo	Information about a live worker instance.
WorkerOptions	Configuration options for creating a Worker instance. Extends QueueOptions.
WorkflowJobDef	-

Type Aliases

Type Alias	Description
BatchProcessor	-
ClosableWorkflow	Returned workflow tree or DAG map plus close() for an injected client.
Field	-
GracefulShutdownHandle	-
Processor	-
ReadFrom	Represents the client's read from strategy.
WorkflowConnection	Connection for helpers. Pass client to keep returned jobs usable.

Variables

Variable	Description
JSON_SERIALIZER	Default JSON serializer used when no custom serializer is provided.
serverlessPool	Module-level singleton for convenient use in serverless handlers.

Functions

Function	Description
chain	Chain: execute jobs sequentially. Each step becomes a child of the next, so step N+1 only runs after step N completes. The last job in the array runs first; the first job in the array runs last and is the top-level parent.
chord	Chord: run a group of jobs in parallel, then execute a callback job with the results. The callback is the parent, the group members are children.
compileSubjectMatcher	Compile an array of subject patterns into a single matcher function. Returns a function that returns true if the subject matches any pattern. Returns null if patterns is empty or undefined (no filtering).
dag	DAG: submit a directed acyclic graph of jobs where each job can depend on multiple other jobs. The graph is validated for cycles and submitted in topological order (leaves first).
gracefulShutdown	Register SIGTERM and SIGINT handlers that gracefully close all provided components. Returns a Promise that resolves when all components have been closed. A second signal while shutdown is still running removes these handlers and re-raises that signal, so the process gets its default handling (exit) instead of the signal being swallowed behind a hung close().
group	Group: execute jobs in parallel. All jobs run concurrently. A synthetic parent job (name: 'group') waits for all children. When complete, the parent's processor receives all children's results via getChildrenValues().
isClusterClient	Detect whether a client is a GlideClusterClient. Uses instanceof with a duck-type fallback for cases where the client comes from a different copy/version of @glidemq/speedkey (dependency duplication).
isTracingEnabled	True when a real OTel API is available (either user-provided or auto-detected).
matchSubject	Match a dot-separated subject against a pattern. - * matches exactly one segment - > matches one or more trailing segments (must be the last token) - Literal tokens match exactly
setTracer	Allow the user to supply their own tracer instance. If not called, the tracer is auto-resolved from @opentelemetry/api.
topoSort	Topological sort of DAG nodes using Kahn's algorithm. Returns nodes in submission order (leaves first, roots last). Throws CycleError if a cycle is detected.
validateDAG	Validate that a set of DAG nodes forms a valid DAG (no cycles). Throws CycleError if a cycle is detected. Throws Error if a node references a non-existent dependency.