Add information about extra debug logging

[cfzhixon]

Summary

Adds information about warp-cli debug extra commands, and about new diagnostic files collected by those commands.

Documentation checklist

• Is there a changelog entry (guidelines)? If you don't add one for something awesome and new (however small) — how will our customers find out? Changelogs are automatically posted to RSS feeds, the Discord, and X.
• The change adheres to the documentation style guide.

[cf-rhett]

/bonk review

[cf-rhett]

Availability version would be good here. Can use 2026.5.0 as the min for now

[ask-bonk]

A few style and consistency suggestions. Please also address the outstanding review comments about adding the availability version (cf-rhett), table rows for TCPIP/WFP traces (cfzhixon), and the Windows ETL pcap note (csujedihy).

[ask-bonk]

Posted a review on PR #30993.

What I found:
3 style/consistency issues in the new "Extra debug logging" section that previous reviewers hadn't caught:

1. Line 156 — "Pcaps" is inconsistent with the table's "Packet capture (PCAP) files"
2. Line 157 — "QLOGs" is inconsistent with the table's "QLog files"
3. Line 159 — The "One-shot connectivity checks" bullet is the only full sentence in the list; rephrased it as a fragment to match the others, and added backticks around warp-diag

I also reminded the author to address the 3 outstanding review comments that haven't been pushed yet: availability version (2026.5.0), missing table rows for TCPIP/WFP traces, and the Windows ETL pcap note.

github run

[cf-rhett]

/bonk review

[ask-bonk]

Thanks for addressing the earlier feedback. The section is well-structured and the content looks accurate. A few remaining items from previous reviewers and minor style nits:

1. cfzhixon's suggestion to add rows in the big table for TCPIP and WFP traces is still outstanding. The current .etl / .xml row groups them together without specifying which trace type each file corresponds to.
2. csujedihy's comment about noting a single ETL for pcaps on Windows could be clearer. The current parenthetical mentions the ETL file but doesn't explicitly state that there is just one.
3. Minor casing inconsistency: .ETL in the bullet should be lowercase .etl to match the table.

[ask-bonk]

Posted a review on PR #30993.

What I found:
3 items remaining from previous reviewers and minor style nits:

1. Line 65 (.pcap row) — With extra debug logging now generating PCAPs automatically, the table description ("manually generated using warp-cli debug pcap commands") is no longer complete. Suggested updating it to mention both manual and automatic generation.

2. Line 66 (.etl / .xml row) — cfzhixon's suggestion to add rows for TCPIP and WFP traces is still outstanding. The current row groups the two file extensions together without specifying which trace type each corresponds to.

3. Line 158 (first bullet) — Two issues: (a) .ETL should be lowercase .etl to match the table entry, and (b) csujedihy's comment about noting there is a single ETL for pcaps on Windows could be clearer — the current text mentions the ETL file but doesn't make the "single file" point explicit. Provided a suggestion addressing both.

github run

[cfzhixon]

/bonk review

[cf-rhett]

/bonk review

[ask-bonk]

Nice work addressing the earlier rounds of feedback. The new Extra debug logging section is well-structured and the table updates cover the new file types well.

A few remaining items to tighten up:

1. .etl vs .xml descriptions overlap. Line 66 says the .etl file contains WFP captures, but line 67 says .xml is also WFP trace files. It's unclear how these relate — is .xml metadata for the WFP capture inside .etl, or a separate trace format? cfzhixon previously noted they would split these up; the current descriptions still group things together in a way that's hard to follow.

2. Terminology mismatch between table and bullet. The *-traceroute-*.txt row (line 121) refers to "edge API endpoint" while the bullet list (line 166) calls it "Device API endpoint." Please use the same name in both places.

3. ETL retention phrasing. Line 159 is clearer than earlier iterations, but the subject of "retains" is still implicit. Consider:

- Packet captures (PCAPs) on all interfaces (on Windows, the original `.etl` file is preserved; normally it is converted to `.pcap` and then removed)

[ask-bonk]

Posted a review on PR #30993.

What I found:
3 remaining items from previous reviewers that haven't been fully resolved:

1. .etl vs .xml overlap (line 66-67) — cfzhixon previously said they would "split them up," but the current .etl row still says it contains "packet captures, netsh TCPIP traces, and WFP captures" while .xml is also described as "WFP trace files." The relationship is unclear.

2. Terminology mismatch (line 121 vs. line 166) — The table row for *-traceroute-*.txt calls it "edge API endpoint" but the bullet list calls it "Device API endpoint."

3. ETL retention phrasing (line 159) — The parenthetical "retains the original .etl file before conversion to .pcap" is slightly ambiguous (what retains it?). Suggested making the subject explicit.

github run

[cf-rhett]

/bonk review

[ask-bonk]

Content looks accurate and previous feedback has been well addressed. I found one remaining style-guide item previous rounds missed: WFP is not spelled out on first use.

[ask-bonk]

Posted a review on PR #30993.

What I found:
After multiple rounds of review, the content is accurate and nearly all prior feedback has been addressed. I caught one remaining style-guide item that previous reviewers missed:

• Line 66 — WFP is used without being spelled out on first use. I suggested expanding it to "Windows Filtering Platform (WFP)" in the .etl table row, which is the first occurrence in the document.

github run

[cf-rhett]

/bonk review

[ask-bonk]

The content is accurate and well-structured after the previous review rounds. I found two remaining style/consistency items in the Extra debug logging section:

1. A minor grammar fix in the .pcap table row.
2. Capitalization and terminology alignment between the *-traceroute-*.txt table row and the bullet list below it.

[ask-bonk]

Minor grammar fix: use via instead of with so the verb carries through:

Suggested change

	| `.pcap` | Packet capture (PCAP) files manually generated using `warp-cli debug pcap` commands, or automatically with [extra debug logging](#extra-debug-logging). |
	| `.pcap` | Packet capture (PCAP) files manually generated using `warp-cli debug pcap` commands, or automatically generated via [extra debug logging](#extra-debug-logging). |

[ask-bonk]

Capitalize Device to match the bullet list (line 166):

Suggested change

	| `*-traceroute-*.txt` | One-shot traceroute to WARP ingress IPs, device API endpoint hostname and IPs, and DoH resolver hostname and IPs; started when [extra debug logging](#extra-debug-logging) is first enabled. |
	| `*-traceroute-*.txt` | One-shot traceroute to WARP ingress IPs, Device API endpoint hostname and IPs, and DoH resolver hostname and IPs; started when [extra debug logging](#extra-debug-logging) is first enabled. |

Also, please align terminology with the bullet list (lines 164–168):

• WARP ingress IPs (table) vs Tunnel edge connection IPs (line 168). Use the same term in both places if they refer to the same thing.
• DoH resolver (table) vs DNS resolver (lines 164–165). Use the same term in both places.

[ask-bonk]

Posted a review on PR #30993.

What I found:
After multiple rounds of thorough review, the content is accurate and nearly all prior feedback has been addressed. I caught two remaining style/consistency issues that previous reviewers missed:

1. Line 65 (.pcap row) — Minor grammar: "or automatically with" should be "or automatically generated via" so the verb carries through.
2. Line 121 (*-traceroute-*.txt row) — Three sub-issues:
   • Capitalization: "device API endpoint" should be "Device API endpoint" to match the bullet list
   • WARP ingress IPs (table) vs Tunnel edge connection IPs (bullet, line 168) — terminology mismatch
   • DoH resolver (table) vs DNS resolver (bullets, lines 164–165) — terminology mismatch

github run

[cf-rhett]

/bonk review

[ask-bonk]

After multiple rounds of thorough review, the content is accurate, well-structured, and all prior feedback appears to be fully addressed.

What was resolved since earlier rounds:

• Version availability (2026.5.0) is noted.
• .etl and .xml table rows clearly distinguish WFP/TCPIP trace content.
• WFP is spelled out on first use in the .etl row.
• Terminology is aligned between the table (*-traceroute-*.txt) and bullet list (Device API endpoint, DNS resolver, Tunnel edge connection IPs).
• The Windows .etl retention note is technically accurate per cfzhixon.
• Grammar in the .pcap table row is correct (“automatically generated via”).

No new MDX syntax or style-guide issues found. LGTM.

[ask-bonk]

Posted an approving review on PR #30993.

What I found after thorough analysis:

All feedback from the previous six review rounds has been fully addressed in the current version of diagnostic-logs.mdx:

• Version availability: "Starting with version 2026.5.0" is correctly noted.
• Table additions: .etl and .xml rows are present and accurately describe WFP/TCPIP trace content.
• Acronym spelling: "Windows Filtering Platform (WFP)" is spelled out on first use.
• Terminology alignment: *-traceroute-*.txt table row and bullet list consistently use Device API endpoint, DNS resolver, and Tunnel edge connection IPs.
• Windows ETL note: The parenthetical on line 159 accurately reflects that extra debug logging preserves the .etl file while other PCAP commands convert and remove it.
• Grammar: The .pcap table row correctly reads "automatically generated via."

No new MDX syntax issues, style-guide violations, or content inaccuracies were found.

github run