Audit and troubleshooting
Find what SOAR did and why: audit events, run statuses, failure reasons, and fixes for flows that don't fire, runs that fail, and dropped trigger events.
Use the run detail and the audit log to see what SOAR did, who asked for it, and why a run ended the way it did.
What's audited
nano writes an audit event for every SOAR change and every run. Find them at Settings → Audit Log, which needs audit:view. See Platform audit.
| Area | Audit actions |
|---|---|
| Settings | soar_settings_updated, soar_runs_canceled_on_disable, soar_notify_settings_updated, soar_notify_test_sent, soar_egress_allowlist_updated |
| Catalog and packs | soar_catalog_imported, soar_catalog_fetch_refused, soar_pack_installed, soar_pack_upgraded, soar_pack_removed, soar_pack_updates_applied |
| Connections | soar_connection_created, soar_connection_updated, soar_connection_removed, soar_connection_tested |
| Flows | soar_flow_saved, soar_flow_archived, soar_flow_drafted, soar_flow_test_requested |
| Runs | soar_run_requested, soar_run_executed, soar_flow_run_requested, soar_flow_run_finished, soar_run_canceled, soar_playbook_automation_queued |
| Approvals | soar_approval_decided |
Audit events never contain credential values. A pivt draft is audited by its size and outcome, not by the text of the request.
Each run also keeps its own record on Automation → Runs: who or what requested it, the flow version, each step's inputs and outputs with secrets redacted, its approvals with the decider and reason, and the redacted vendor request.
Run statuses
| Status | Meaning |
|---|---|
queued | Waiting for a worker, a free request slot, or a retry delay |
running | A worker is executing a step |
suspended | Waiting for an approval |
succeeded | Every step finished |
failed | A step failed, or an approval expired. The failure reason says why. |
canceled | Someone canceled the run, an approval was rejected, SOAR was turned off, or the run's authority no longer holds |
skipped | The trigger was recorded but not run. The runs list and the run trace give the reason: rate limited (the flow's run cap), connection missing, connection ambiguous, pack not installed, flow invalid, schedule invalid or not admitted. Skipped runs are kept 7 days. |
Accepted or queued doesn't mean the action happened. Check the step result.
Failure reasons
The run detail shows the failure reason. Steps record the same reasons.
Vendor and action outcomes
| Reason | Meaning | What to do |
|---|---|---|
auth_failed | The vendor refused the credentials or their permissions (HTTP 401 or 403). No change was applied. | Fix the credentials or the account's permissions in the vendor, edit the connection, and click Test connection. |
rate_limited | The vendor returned HTTP 429, or the pack's request budget for the minute or day is used up | Wait, or reduce how often the flow runs. |
ambiguous | The request may have reached the vendor, but nano didn't get a clear answer | Check the target in the vendor. nano resends only actions that are safe to retry; otherwise a person must approve a resend. |
egress_unavailable | nano couldn't open a connection to the vendor under the outbound rules | Check DNS, your firewall, and SOAR_EGRESS_PROXY_URL if you use a proxy. |
invalid_config | The connection's settings are invalid for this action | Edit the connection. |
unsupported | The vendor doesn't support this request | Check the pack's documentation. |
action_failed | The vendor returned another error | Open the step to see the redacted request and status. |
step_attempts_exhausted | The step reached 32 attempts | Fix the cause and start a new run. |
Flow and connection changes
| Reason | Meaning |
|---|---|
pack_version_changed | The pack was upgraded after the run started. Start the run again. |
pack_unavailable, action_unavailable | The pack or action was removed or withdrawn. |
connection_changed | The connection was edited after the run started. Start the run again. |
connection_removed | The connection was deleted. |
credentials_unavailable, credentials_invalid | nano couldn't open the stored credentials, or they don't match what the pack needs. Re-enter them. |
secret_in_inputs | A step input contained a connection credential, so nano didn't send it. |
action_input_unresolved, action_input_invalid | An input reference had no value, or the value didn't match what the action accepts. |
flow_invalid, flow_evaluation, invalid_run, unsupported_module | The flow no longer compiles, or an expression couldn't be evaluated. Validate the flow. |
Case steps
Case steps have their own reasons, such as case_missing, case_source_scope_refused and case_status_transition_invalid. See Case steps.
Searches, Judge, and notifications
| Reason | Meaning |
|---|---|
search_query_invalid | The query didn't parse, or named a field that doesn't exist |
search_timeout | The search took longer than 30 seconds |
search_too_large, search_admission_unavailable, search_execution_failed, search_unavailable | The search couldn't run or return |
search_binding_invalid, search_bounds_invalid | A bound value was missing or unusable, or the window or row limit is out of range |
judge_budget_exhausted | The daily token budget is used up. See AI Judge. |
judge_timeout, judge_provider_error, judge_unavailable | The AI provider didn't answer in 60 seconds, returned an error, or isn't configured |
judge_binding_invalid | An evidence value was missing or too large |
notification_cap_reached | 200 notifications were already sent in the last 24 hours |
notification_destination_unknown | The recipient or webhook endpoint isn't configured in Settings → SOAR |
notification_changed | The content or destination changed after it was approved |
notification_render_invalid, secret_in_notification | The message couldn't be rendered, or still contained a secret after redaction |
smtp_not_configured, smtp_credential_too_short, smtp_auth_failed, smtp_connect_failed, smtp_tls_failed | The SMTP relay isn't set up or refused the message. Check Settings → SOAR and use the test send. |
Approvals and authority
| Reason | Status | Meaning |
|---|---|---|
approval_rejected | canceled | An approver rejected the step |
approval_expired | failed | Nobody decided before the approval expired |
approval_authority_revoked | canceled | The approver lost soar:approve or access to the run's data before the step ran |
authority_revoked | canceled | The person the run acts for lost a permission the run needs, or can no longer act |
authority_missing | canceled | The run has no readable record of who it acts for |
source_visibility_revoked | canceled | The person the run acts for can no longer see the triggering alert, its data, or the linked case |
playbook_acl_revoked | canceled | The person the run acts for can no longer run the playbook |
canceled | canceled | Someone canceled the run |
soar_disabled | canceled | SOAR was turned off |
Troubleshoot
Need help with these checks? Contact support with the run ID and the time with timezone.
A live flow doesn't fire on an alert
Check, in order:
- The flow is
live, notdraft. Only live flows fire. - SOAR is on at Settings → SOAR.
- The trigger is Alert committed, and its event match, severity, and rule filters fit the alert. Test the flow with that alert to see whether it compiles for the alert trigger.
- The person who last saved the flow can see the alert, including its log source. nano skips a flow for an alert its owner can't see.
- The flow still compiles. A pack that was removed, or a Pack default connection on a pack that now has two connections, stops the flow from starting runs.
- The flow hasn't already run for this alert. Each alert starts a flow once.
A scheduled flow went back to draft
nano returns a live schedule flow to draft when it can no longer start a run, for example because it no longer compiles or its owner lost a permission. Fix the cause, save, and activate it again.
The Flows page says trigger events were dropped
Trigger events dropped means the trigger queue held 20,000 waiting events and nano dropped new ones. Alerts and cases were still saved, but those events didn't start runs. Narrow your triggers with severity and rule filters and event matches, then run the affected flows manually for the alerts you still need.
A trigger event was dead-lettered
When nano fails to start runs for an event 10 times in a row, it sets the event aside and logs a warning. Dead-lettered events don't start runs. nano keeps them for 7 days, up to the newest 10,000. Fix the cause, then run the flow manually for that alert or case.
A run waits and never finishes
Open the run. suspended means it is waiting for an approval: check Automation → Approvals. The person who started the run, or who last saved a triggered flow, can't approve it; someone else with soar:approve must. If nobody decides, the approval expires and the run fails.
queued for a long time means all workers or request slots are busy. See How runs behave under load.
An action can't be approved
The approver needs soar:approve and an interactive sign-in, can't be the requester, and must be able to see all of the step's data. If source visibility hides any of it, the approver can reject the step but not approve it.
A step says the executor restarted
If a jobs service process stops mid-step, nano reclaims the run after its lease expires, about 16 minutes later. A step that may have reached the vendor is marked ambiguous. A flow run goes back to the queue; a single action fails with ambiguous so a person can check the target first.