13% of Part I
4 sub-topics · this deck covers sub-topic 3 of 4
↑ Module OverviewPreserving execution records lets you: identify and diagnose error conditions, retrace executed flows to find bottlenecks and optimization targets, and facilitate auditing for compliance to internal and external policy.
Flow input and output (IO) data — every card has configurable settings plus input/output values. Opting to Save flow data lets Okta record the specific data passing through each card during execution, so you can review every card's values afterward.
Flow metadata — Okta records metadata for every execution event regardless of the Save flow data setting, including the type of flow, the execution timestamp, and the final execution status.
Recorded data lives on the Execution History page of a flow. "Anyone with the appropriate privileges can view these details for up to 30 days." This 30-day retention window is the same limit that appears again under System Limitations, listed there as the "Data time to live" for execution history.
The Details column of the Execution History pane assigns each execution one of three icons:
Low-latency (lightning bolt) — every card in the execution ran efficiently, so the flow ran in low-latency mode. This is common for latency-sensitive webhook and API Endpoint use cases.
Standard (dash) — the execution ran at a standard rate because it contains a card ineligible for low-latency mode.
Throttled (exclamation mark) — the execution experienced higher latency from resource constraints. Once throttled, a flow remains in throttled mode for 15 minutes and isn't eligible for low-latency mode during that time.
Okta describes four common scenarios you'll see across a flow's execution list:
All standard — every execution contains a card that doesn't qualify for low-latency. All low-latency — every execution ran efficiently. Mixed — only some executions were low-latency; check the flow for cards you can modify to remove the latency issue. All throttled — every execution exceeded resource limits.
Execution History is the diagnostic layer that sits directly beneath everything else in this section: it's how you'd confirm whether the API Endpoint flow from the previous sub-topic ran in low-latency mode, and it's the tool you'd reach for when a flow starts hitting the rate limits and execution limits covered next under System Limitations.
Watch for this in your own sandbox run:
→ Assuming flow metadata (type, timestamp, final status) requires the Save flow data option — it's recorded regardless. Only the per-card IO data depends on that setting.
→ Treating a single "Standard" execution as a sign of a persistent problem — a lone card ineligible for low-latency is enough to mark that one execution Standard, without any resource constraint being involved.
Sources: help.okta.com — ext-about-save-data-flow-history (Flow execution history), ext-flow-execution-details (Execution history details)
Next: Sub-topic 4 →