feat: Unified environment log reader (plugin traces, async jobs, workflows, flow runs)

[TomProkop]

Context

During a conference demo session building a Warehouse Management app end-to-end, plugin steps were imported as disabled and plugins did not fire. There was no CLI way to diagnose this — we had to raw-query sdkmessageprocessingstep via SQL to discover statecode=1. If plugins had thrown runtime errors, there would have been no way to see trace logs without opening the browser.

Currently, debugging Dataverse runtime behavior requires the maker portal or manual OData/SQL queries against system tables. The CLI should provide a unified log surface for runtime diagnostics.

Note: Deployment/import logs are already covered by txc environment deployment list/show. This issue is specifically about runtime execution logs that have no CLI support today.

Proposal

Add a txc environment log command family for runtime diagnostics:

Command	Source Table / API	Use Case
txc environment log list	Multiple (unified)	Recent logs across all sources, sorted by time
txc environment log plugin-trace	plugintracelog	Plugin execution traces, errors, custom Trace() output
txc environment log async-jobs	asyncoperation	Background job status — workflows, async plugins, bulk ops
txc environment log workflow	workflowlog + asyncoperation	Classic workflow run history with step details
txc environment log flow-runs	Power Automate API	Cloud flow run history, trigger details, action failures

Key flags

--since 1h              # Time window (1h, 30m, 7d)
--entity <name>         # Filter to logs related to a specific entity
--plugin <name>         # Filter by plugin class name
--errors-only           # Only show failures/exceptions
--follow / --watch      # Tail mode — stream new logs as they appear
--correlation-id <guid> # Trace a single request across plugin chain

MCP tool equivalents

Tool	Parameters
environment_log_list	since, entity, errors-only
environment_log_plugin_trace	since, plugin, entity, errors-only, correlation-id
environment_log_async_jobs	since, entity, status-filter
environment_log_flow_runs	since, flow-name, status-filter

Example workflows

# "Why didn't my plugin fire?"
txc environment log plugin-trace --since 10m --errors-only

# "What happened to my async workflow?"
txc environment log async-jobs --since 1h --entity udpp_warehousetransaction

# "Show me everything that happened in the last 5 minutes"
txc environment log list --since 5m

# "Tail logs while I test in the browser"
txc environment log plugin-trace --follow

Implementation notes

• plugin-trace queries plugintracelog entity (columns: typename, messagename, exceptiondetails, messageblock, performanceexecutionstarttime, operationtype)
• async-jobs queries asyncoperation (filter by operationtype, regardingobjectid, statuscode)
• flow-runs requires Power Automate Management connector or direct API calls to flow.microsoft.com
• --follow can poll with configurable interval (default 5s)
• Consider --json output mode for piping to jq
• Out of scope: deployment/import logs (already covered by txc environment deployment list/show)

Priority: 🔴 High | Effort: Large