Deployment and upgradesOfficial documentation7 min readDeployment and upgrades

Resolve Stalled or Slow Dataset Indexing

Dataset indexing issues, including stalled progress or slow performance, are common during FastGPT self-hosted deployments.

Dataset indexing issues, including stalled progress or slow performance, are common during FastGPT self-hosted deployments. The first critical troubleshooting step is to review application log error messages to identify the specific failure mode, as this directly points to the root cause of the problem.

Standardized Troubleshooting Symptom Matrix

Use the following matrix to map observed indexing issues to their confirmed root causes, based on official FastGPT diagnostic data:

Observed SymptomConfirmed Root Cause
Successful external service connectivity verification, but no indexing progress is recordedThe vectorModels configuration field is not properly configured for the deployment
Failed connectivity verification attempts, and no indexing workflow initiates at allAPI call failure; the deployment cannot establish a connection to OneAPI or OpenAI services
Visible incremental indexing progress, but task completion occurs at an extremely slow rateSuboptimal API key configuration or account rate limitations on the connected LLM/vector service

Resolving Stalled Indexing with Successful Verification

When the system can confirm connectivity to required services but indexing does not advance, the only required fix is to validate the vectorModels configuration. This field must be populated with the correct vector model details for the deployment; missing or empty configuration will block indexing entirely even if basic connectivity checks pass.

Fixing Complete Indexing Failure

If both connectivity verification and indexing fail entirely, the root cause is a failed API connection. Confirm that the deployment has valid access to OneAPI or OpenAI endpoints, and that all authentication credentials (including API keys) are correctly entered. Any failed API calls during pre-deployment checks will prevent both verification and indexing from starting.

Addressing Slow Indexing with Visible Progress

For deployments where indexing progresses but at an unacceptably slow pace, the primary issue is often related to API key quality or account tier restrictions. Free-tier OpenAI accounts have strict rate limits, including per-minute request caps and daily request limits that will throttle indexing speed. Switching to a paid-tier API key or upgrading the associated account will resolve this slowdown.

Source: FastGPT official source

Applicability and version scope

Use this page for the documented Deployment and upgrades scenario. Confirm the FastGPT, dependency, API, and deployment versions in the official source before applying a change.

Safety guardrails

Use [REDACTED_CREDENTIAL] for credentials and private data. Confirm the documented environment and version before review.

Rollback guidance

Restore the prior technical-content authority snapshot. Restore saved configuration and data snapshots, then repeat the smallest verification scenario.