The identity and limits of one run. RunConfig() is evaluated per call, so each gets a fresh RunId. The tenant is part of the thread's identity: it is recorded on every checkpoint, and admission refuses a run whose tenantId differs from the latest checkpoint's (GraphError.TenantMismatch; None and Some differ). The principal is recorded on the run's RunStarted, RunRecovered and RunResumed events and is never checked.
metadata, when not empty, is the thread's ThreadMetadata (design 4.52): a top-level start sets it in its claim commit, replacing what the thread had, and refuses metadata that breaks its limits with a ValidationError before it claims anything. recover, resume and a child thread's run leave it as it is; change it with GraphRuntime.updateMetadata.
interruptBefore and interruptAfter are the run's static breakpoints (BreakpointPhase): a task of a node in interruptBefore is held before it runs, and one of a node in interruptAfter after it ran, each as an interrupt of its own that GraphRuntime.resume answers with Breakpoint.proceed. They belong to the run, like its budgets: a resume or recover brings its own, and a continuation already past a breakpoint is not held by it again. Admission refuses a node the graph does not have with a ValidationError.
Attributes
- Companion
- object
- Source
- RunConfig.scala
- Graph
-
- Supertypes
Members list
Value members
Concrete methods
Attributes
- Source
- RunConfig.scala
How many settled (Completed or Failed) child threads of the run's thread to keep (design 4.41, default RunConfig.DefaultChildRetention): once the run's terminal checkpoint commits, and at a later start on the settled thread, older settled children are deleted, each with everything below it. At that start a child an earlier run left Running or Suspended can never be reattached, so it is first settled RunOutcome.Failed("abandoned", ...) and then counts; otherwise one is never deleted, and nothing is at recover or resume. Throws IllegalArgumentException below 0.
How many settled (Completed or Failed) child threads of the run's thread to keep (design 4.41, default RunConfig.DefaultChildRetention): once the run's terminal checkpoint commits, and at a later start on the settled thread, older settled children are deleted, each with everything below it. At that start a child an earlier run left Running or Suspended can never be reattached, so it is first settled RunOutcome.Failed("abandoned", ...) and then counts; otherwise one is never deleted, and nothing is at recover or resume. Throws IllegalArgumentException below 0.
Attributes
- Source
- RunConfig.scala
Holds every task of these nodes after it ran, replacing the nodes set before.
Holds every task of these nodes after it ran, replacing the nodes set before.
Attributes
- Source
- RunConfig.scala
Holds every task of these nodes before it runs, replacing the nodes set before.
Holds every task of these nodes before it runs, replacing the nodes set before.
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
Prunes the thread to policy (Checkpointer.prune) as the run is admitted: once the store has granted its claim and before the run commits anything, so a run refused at admission - a blank query, a busy thread, a pending interrupt, another tenant's thread - leaves the thread unchanged, nothing else writes to the thread while it prunes, and nothing of the run exists yet to be pruned, by any limit, so its own replay is whole. Pruning never removes the latest checkpoint or its pending writes, so what the run continues from is unchanged; earlier runs' checkpoints and events beyond the policy are removed, raising the replay floor for their handles. RetentionPolicy.keepAll removes nothing and is skipped. A prune that fails is logged at WARN and the run goes on: retention is housekeeping.
Prunes the thread to policy (Checkpointer.prune) as the run is admitted: once the store has granted its claim and before the run commits anything, so a run refused at admission - a blank query, a busy thread, a pending interrupt, another tenant's thread - leaves the thread unchanged, nothing else writes to the thread while it prunes, and nothing of the run exists yet to be pruned, by any limit, so its own replay is whole. Pruning never removes the latest checkpoint or its pending writes, so what the run continues from is unchanged; earlier runs' checkpoints and events beyond the policy are removed, raising the replay floor for their handles. RetentionPolicy.keepAll removes nothing and is skipped. A prune that fails is logged at WARN and the run goes on: retention is housekeeping.
A child thread's run (design 4.42) is the exception: its admission prunes only after the claim commit, which stores the usage folded from the earlier attempts' terminal events, and before the run is launched. A run refused at admission, or whose claim commit fails, still leaves the thread unchanged; but the claim's own checkpoint and event exist by then and count against the limits (maxEvents is at least 1 on a child), and maxAge may remove that event, raising the replay floor above the run's start.
Attributes
- Source
- RunConfig.scala
withRetention, or no pruning for None.
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
The trace context this run continues (design 4.58): the run's span takes it as its remote parent, so the run is in that context's trace - a submitter's request, a queue entry's recorded context, the original run's context for a recovery on another machine. None, the default, starts a trace of the run's own, derived from its RunId. A delegated child's run is given the context its origin recorded (RunOrigin.trace), whatever is set here. Tracing reads it; nothing else in the run does.
The trace context this run continues (design 4.58): the run's span takes it as its remote parent, so the run is in that context's trace - a submitter's request, a queue entry's recorded context, the original run's context for a recovery on another machine. None, the default, starts a trace of the run's own, derived from its RunId. A delegated child's run is given the context its origin recorded (RunOrigin.trace), whatever is set here. Tracing reads it; nothing else in the run does.
Attributes
- Source
- RunConfig.scala
Attributes
- Source
- RunConfig.scala
