Skip to main content
Control flow nodes orchestrate how and when parts of the graph execute—branching on data, isolating failures, looping over sets, fanning out partitions, and kicking off other pipelines.

​
Conditional

Conditional implements if / else branching based on expressions or parameter values. Configuration:
  • Predicate: Boolean expression or rule set evaluated at the decision point.
  • True / false paths: Distinct downstream subgraphs wired from each branch output.
Typical use: When load_type = 'full_refresh', run a truncate-and-load subgraph; otherwise run incremental merge logic.
Clearly label branch outputs in large graphs so reviewers know which path corresponds to which business case.

​
Error Handler

Error Handler wraps a try / catch region around a subgraph. Configuration:
  • Protected region: Nodes whose failures should not abort the entire run (when configured).
  • Catch path: Steps for logging, notifications, dead-letter writes, or compensating actions.
  • Retry policy (when available): Backoff and max attempts for transient failures.
Typical use: Isolate calls to flaky partner APIs—on failure, write the payload to a quarantine table and alert on-call.

​
ForEach

ForEach iterates over records or a list parameter, executing a child graph per item. Configuration:
  • Iterator source: Query, parameter list, or column collection.
  • Loop body: Downstream nodes referencing the current item via variables.
  • Parallelism: Limit concurrent iterations to protect shared systems.
Typical use: Process ten regional exports with the same logic—parameterize bucket names per row instead of cloning pipelines.

​
Split

Split partitions a dataset into multiple outputs by predicate, hash, or key ranges—useful for parallel writes or A/B processing paths. Configuration:
  • Rules per output: Mutually exclusive filters or modulo buckets.
  • Default branch: Catch-all for unexpected values.
Typical use: Send country = 'US' rows to one warehouse share and EU rows to another while keeping one authoring graph.

​
Trigger Pipeline

Trigger Pipeline starts another Planasonix pipeline by reference. Configuration:
  • Target pipeline and version or label (per product).
  • Parameters / variables passed into the child run.
  • Wait behavior: Fire-and-forget vs wait for completion for chaining SLAs.
Typical use: A lightweight orchestrator graph that sequences domain-specific jobs after a shared staging refresh completes.

​
Trigger Reverse ETL

Trigger Reverse ETL kicks off a reverse ETL sync or related outbound job defined in your workspace. Configuration:
  • Sync identifier: Which reverse ETL configuration to run.
  • Rowset or keys (when supported): Limit the sync to a cohort produced upstream.
Typical use: After curated metrics land in the warehouse, push refreshed segments to Salesforce or an ads platform on demand.
Chained triggers can amplify load. Rate-limit downstream SaaS APIs and respect partner daily quotas.

​
Modeling patterns

Keep “traffic cop” pipelines thin—Trigger Pipeline nodes and parameters—while heavy transforms live in dedicated graphs you can test independently.
Pair Error Handler catch paths with Notification nodes so failures become tickets, not silent skips.

Orchestration

Schedules, triggers, and platform-level execution.

Pipeline canvas

Wire control-flow outputs like any other node.