Overview

‍

The Advanced Settings panel gives workspace administrators and power users precise control over TaskForge's runtime behavior, performance characteristics, and integration configuration. This section covers the most commonly used advanced configuration options and when to apply them.

‍

‍

API rate limit configuration

‍

By default, TaskForge enforces per-workspace API rate limits to ensure fair usage across all customers. On paid plans, these limits can be customized to match your throughput requirements.

‍

To configure rate limits:

‍

1. Go to Settings → Advanced → API Rate Limits.
2. Adjust the requests-per-second (RPS) slider for each API category (Pipeline, Model, Analytics, Webhooks).
3. Set burst allowance to handle traffic spikes without triggering throttling.
4. Click Save to apply changes immediately.
5. Custom rate limits take effect within 60 seconds of saving. If you require limits beyond the plan maximum, contact the TaskForge team to discuss an Enterprise arrangement.

‍

‍

Model inference parameters

‍

For workspaces using AI model integration, the Advanced Settings panel exposes low-level inference parameters that control how models process requests:

‍

• Timeout: Maximum allowed inference duration in milliseconds before the request is cancelled. Default is 30,000ms.
• Retry policy: Number of automatic retries on transient errors, with configurable backoff strategy.
• Concurrency limit: Maximum number of simultaneous inference requests per model.
• Warm-up mode: Keep model instances pre-loaded in memory to eliminate cold-start latency.
• Output caching: Cache identical inference inputs to reduce redundant model calls and lower costs.

‍

Custom pipeline execution environment

‍

By default, pipeline tasks run in a shared TaskForge managed execution environment. For tasks with specialized requirements, you can configure a custom execution environment:

‍

1. Navigate to Settings → Advanced → Execution Environments.
2. Click Create Environment and select the base image (Node.js, Python, or custom Docker image).
3. Define environment variables, memory allocation, and CPU limits for this environment.
4. Assign the environment to specific pipelines from the pipeline configuration page.

‍

Custom environments are built and cached on first use. Subsequent deployments reuse the cached image unless the base configuration changes.

‍

‍

Workspace data export

‍

You can export a full snapshot of your workspace data at any time from Settings → Advanced → Data Export. The export includes:

‍

• All pipeline and task configurations in JSON format
• Execution history and logs for the selected date range
• Model deployment records and version history
• Workspace member list and permission assignments
• API key metadata (keys themselves are not included for security reasons)

‍

Exports are generated asynchronously and delivered to your registered email address as a secure download link. Large workspaces may take up to 30 minutes to compile.

‍

‍

Debug mode

‍

Debug mode enables verbose logging across all pipeline executions, capturing full request and response payloads, step timing breakdowns, and internal task state transitions. This is useful for diagnosing complex issues but should not be left enabled in production due to the performance overhead and increased log volume.

‍

To enable debug mode for a specific pipeline:

‍

1. Open the pipeline configuration.
2. Click Advanced Options.
3. Toggle Enable Debug Mode.
4. Run the pipeline — all debug output will appear in the Log Explorer tagged with [DEBUG].

‍

Notification preferences

‍

Advanced notification settings allow you to configure exactly which events trigger alerts and through which channels. You can set different notification rules per pipeline, per team, or workspace-wide. Supported channels include email, Slack, Microsoft Teams, PagerDuty, and any custom webhook endpoint. Notification rules support condition logic so you can suppress low-severity alerts during off-hours or route critical failures to an on-call escalation path.