Troubleshooting

Find the fastest path to a fix when something in Jutsu does not behave as expected. This hub groups the most common problems by area; each table maps a symptom to its likely cause and the action that resolves it, with links to the in-depth page for that area.

Ingestion

No events or alerts are reaching Jutsu, or they stopped arriving. Most ingestion problems come down to the forwarder, the ingest key, or the endpoint.

SymptomLikely causeFix
No new events or alerts appear at allThe forwarder or connector is not runningConfirm your Wazuh integration, syslog forwarder, or connector is up and shipping data. See Wazuh data source.
Requests are rejected with 401The ingest API key is missing, invalid, expired, or inactiveVerify the key is active and sent in the X-API-Key header. See Ingestion API.
Some events index, others are droppedPer-item payload validation failed on a batchRead the validationErrors array in the response, fix the bad entries, and resend. See Ingestion API.
Events go nowhere with no error in JutsuForwarder points at the wrong webhook endpointConfirm the forwarder targets the correct ingest webhook URL (Wazuh vs. syslog). See Ingestion API.

Credentials

An AgentSOAR provider credential fails to validate or stops working. Validation almost always points to a secret, a permission, or a missing consent.

SymptomLikely causeFix
Credential validation fails on connectWrong client secret, key, or tokenRe-enter the secret from the provider and revalidate. See Credentials & domains.
Validation passes but actions are deniedThe credential lacks the required scopes or rolesGrant the permissions the connector needs on the provider side. See the connector page, e.g. Microsoft 365.
Microsoft consent error during setupAdmin consent was never granted for the appHave a tenant admin approve admin consent, then revalidate. See Microsoft 365.
A working credential suddenly failsThe secret rotated or expiredRe-authenticate the credential. See Credentials & domains.

Response actions

An AgentSOAR action fails, or you need to undo one. Every execution records a status and, on failure, a reason that isolates the cause.

SymptomLikely causeFix
Execution failed with credential_invalidThe provider credential was rejectedRotate or re-authenticate the credential, then re-run. See Executions & revert.
Execution failed with cloud_api_errorThe provider's API returned an errorCheck provider-side status and limits, then retry the action. See Executions & revert.
Execution failed with inventory_not_foundThe target host or user could not be resolvedConfirm the asset exists in inventory and the identifier is correct. See Executions & revert.
Execution failed with input_invalidThe action inputs failed validationCorrect the inputs (IP, email, domain, reason) and re-run. See Executions & revert.
A run is stuck awaiting_reauthThe credential is invalid or expired mid-runRe-authenticate within the cutoff window before it becomes expired_awaiting_reauth. See Credentials & domains.
Revert did not restore the changeThe action was not revertible, or revert_failedRetry the revert; note AWS isolate is not revertible once the host was already isolated. See Executions & revert.

Ingest API errors

A request to the ingest API returns an error status. The status code tells you whether the problem is the API key, access, or a rate limit.

SymptomLikely causeFix
401 UnauthorizedThe ingest API key is missing, invalid, expired, or inactiveSend a valid, active ingest API key in the X-API-Key header. See Ingestion API.
403 ForbiddenThe credential is valid but lacks access to the resourceUse an account or key with the right access for the action. See Role-based access control.
429 Too Many RequestsA rate limit was exceededBack off and retry, respecting the rate-limit response headers.

Not every status applies to every endpoint. Treat these as the general model and confirm per-endpoint behavior against your deployment.

Still stuck?

If an issue is account-specific or you cannot resolve it from these tables, contact your Jutsu administrator or book a call. See Support for all the ways to get help.