idmp-workflow-datasource-diagnose

name	idmp-workflow-datasource-diagnose
description	IDMP datasource diagnosis workflow. Read the connection, probe connectivity, inspect databases and metadata, compare model mappings, and reread after every probe or write.
metadata	{"author":"TaosData","version":"0.6.0","owner_team":"AI","requires":{"bins":["idmp-cli"]},"cliHelp":"idmp-cli datasource --help"}

workflow: datasource diagnose

Read ../idmp-shared/SKILL.md first.

Recommended references

• references/datasource-diagnose.md
• ../idmp-datasource/SKILL.md

Missing context to resolve first

• Credential source.
• Verification rereads.
• The target connection, database, or table that the operator thinks is broken.
• Whether the complaint is reachability, metadata visibility, or model mismatch.
• Which element or template mapping should be compared back to source metadata.

Constrained live behaviors

• Do not copy connections get into connectivity create.
• datasource check list for the built-in TDengine listener.
• After any write-like probe or import, reread.
• Read the connection object before probing reachability.
• Additional properties and metadata are evidence; they do not replace a live connectivity probe.
• Metadata diagnosis stays ordered: connection -> probe -> db -> table -> column -> model mapping.

Execution flow

1. Read idmp-cli datasource connections list and idmp-cli datasource connections get --params before any probe.
2. Use idmp-cli datasource additional-properties list-get --params to capture datasource-side context.
3. Reread idmp-cli datasource check list and then run idmp-cli datasource connectivity create --ack-risk --data to separate summary health from live reachability.
4. Use idmp-cli datasource dbnames list --params, idmp-cli datasource tablenames list --params, and idmp-cli datasource columninfo create --ack-risk --params to lock the metadata boundary.
5. Compare that metadata with idmp-cli attr-template elements attributes --params, and use idmp-cli datasource csv create --ack-risk --params only when the diagnosis includes a write-like CSV probe.
6. Reread idmp-cli data records list if the diagnosis also touches import or export traces.

Exception paths

• If the connectivity probe fails, fall back to datasource check list plus metadata reads before interpreting missing tables as schema bugs.
• If metadata exists but the model does not match, report the exact column or mapping mismatch.
• If record traces are absent, say whether the gap is ingestion history or datasource visibility. Capture whether the root cause is connectivity.

Validation scenarios

1. Listener health confirmation

Start with idmp-cli datasource connections list and idmp-cli datasource connections get --params. Only then run idmp-cli datasource check list.

2. Redacted-payload connectivity failure isolation

Use idmp-cli datasource connectivity create --ack-risk --data after a redaction-aware connection read. Reachability failures should not trigger secret hunting.

3. Metadata versus model mismatch

Pair idmp-cli datasource columninfo create --ack-risk --params with idmp-cli attr-template elements attributes --params. The result should name the exact mapping mismatch.

4. Table discovery failure

Walk idmp-cli datasource dbnames list --params and idmp-cli datasource tablenames list --params in order. The failing step should become the reported boundary.

5. CSV import trace during diagnosis

If the datasource complaint touches ingestion, reread idmp-cli data records list. Keep record-trace evidence separate from connectivity evidence.