Logs and Traces

Logs and traces give you visibility into what happens when your flows process data. They are accessed from the same section of the navigation and are linked together by trace GUIDs.

• Logs record events that occur during processing: informational messages, warnings, and errors.
• Traces tie logs and result payloads together into a single processing run, visualized as a flow chart. Per-node payloads (the data each node emitted) live inside the trace, alongside the graph.

Logs

Logs capture processing events from nodes and services as your flows run. Each log entry includes a message, a severity level, and links to the flow, node, and trace that produced it.

Log Levels

• Info: routine processing events (e.g., file fetched, document converted)
• Warn: non-fatal issues that may need attention
• Error: failures that prevented a node from completing
• Debug: detailed diagnostic information

Filtering and Searching

The Logs page lists the newest logs first, under a heading for each day. Times show in your browser’s time zone, which is named at the top of the page.

Use the controls at the top of the Logs page to narrow down what you see:

• Level tabs: filter to a specific severity (All, Info, Warnings, Errors, Debug)
• Flow: filter to logs from a specific flow
• Service: filter to logs from a specific node service (e.g., SftpFetchService, EdiToJsonService)
• Date: limit the list to the last hour, 24 hours, 7 days, 30 days, or 45 days
• Search: find logs by trace GUID or log ID

Active filters appear as badges below the controls. Click the X on a badge to remove that filter.

Under each message are the flow name and the first characters of the trace ID. Click the trace ID to open that run in the Trace Viewer.

Log Details

Click a log to open its detail panel. The panel shows:

• The level, service, time, and full message
• Data: structured data the log carries, when it has any
• Source: the flow and node that wrote the log, the trace ID, and the log ID
• In this trace: every log from the same run, oldest first, with the current one highlighted

The arrows at the top step to the earlier or later log in the same trace.

When Logs Appear

Logs are written in batches, so a new log can take up to a minute to appear.

Trace Viewer

A trace represents a single processing run through one or more flows, identified by a trace GUID. The trace viewer assembles all logs and per-node payloads for that trace into a visual flow chart so you can see exactly what happened at each step.

Accessing the Trace Viewer

From the Logs page, click the magnifying glass icon next to any trace GUID. You can also click Trace Viewer from a log detail panel, or open the Trace Viewer directly and paste a trace GUID.

Flow Chart

The trace viewer displays the flow’s nodes in a left-to-right diagram. Each node card shows:

• The node kind (input, transformation, splitter, branch, sink, or output) and service
• The number of results produced
• A green border when all results succeeded
• An amber border when a result raised a warning and none errored
• A red border when any result contains an error
• A dashed border when the node produced no results (it was not reached during this run)

Node Details

Click a node in the flow chart to view its details below. The detail panel has three sections:

1. Input: payloads from upstream nodes that fed into this node. Entry nodes (with no upstream connection) show “No incoming result.”
2. Output: payloads produced by this node. Each card shows the result ID, timestamp, metadata, and data payload, with a format tag (EDI, JSON, XML) so X12 envelopes render as raw text rather than as escaped JSON.
3. Logs: log entries specific to this node.

Cards are expanded by default. Click the Data toggle to collapse or expand the payload view. Use the copy buttons to grab the metadata or data payload to your clipboard.

The same payloads can be read through the API: GET /platform/results/:id/payload returns one result’s payload, and the MCP result_payload tool does the same for a coding agent. The trace endpoint lists each result’s payload format and size. See Results and Artifacts.

Diagnosing Errors

When a trace contains errors, a Diagnose Errors button appears in the trace viewer header. This opens an AI diagnostics panel that analyzes the error results and logs across the node graph. The diagnosis also reads the configuration the trace ran against: the flow, the partner and its connection, and the mapping, implementation and source behind a transformation. It names what it read, and it recommends the fix where the fault is, whether that is your integrated system, your Tediware configuration, or the trading partner.

Using Diagnostics

1. Click Diagnose Errors to open the panel
2. Optionally type additional context about the errors (e.g., what changed recently, what the expected behavior is)
3. Click Diagnose to generate an analysis
4. After the initial diagnosis, use the chat input to ask follow-up questions

You can choose which AI model to use from the dropdown in the panel header. Use Re-diagnose to start a fresh analysis, or Clear History to reset the conversation.

Data Retention

Logs and per-node payloads are automatically removed after 45 days. Payloads linked to billing records are kept until those records are also removed. EDI transaction records are hidden after 45 days and permanently removed after 120 days. Retrieve your data via webhooks, the API, or direct downloads before the retention window closes.