Add isolate reason, isolated link count, and -nz filtering for fabric cli commands

[saksarav-nokia]

What I did

Made the following changes in fabric cli commands . The corresponding sonic-swss PR sonic-net/sonic-swss#4704
Display ISOLATE_REASON from STATE_DB in fabric isolation output.
show isolated_links in fabric capacity alongside operating/total links, matching new FabricPortsOrch telemetry.
Add -nz/--nonzero to fabricstat and show fabric for port, queue, rate, and isolation views to hide rows with zero counter deltas or rates.
Extend fabricstat unit tests and mock STATE_DB/COUNTERS_DB fixtures for the new fields and nonzero filtering.

sonic-mgmt PR: sonic-net/sonic-mgmt#25637

How I did it

Modified "show fabric isolation" command out

Modified "show fabric monitor capacity" output.

Added -nz or --nonzero option to all "show fabric" commands

How to verify it

Added unit tests
Also modifying existing sonic-mgmt tests.

Previous command output (if the output of a command-line utility has changed)

New command output (if the output of a command-line utility has changed)

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[saksarav-nokia]

@vmittal-msft @ysmanman @kenneth-arista , Please review.
This is a minor enhancement to the fabric isolation and fabric monitor feature for better debugging. With these changes, the "show fabric monitor capacity" and 'show fabric isolation' commands tell us what is exact reason for the isolated ports Instead of checking multiple cli commands.

[Copilot]

Pull request overview

This PR updates SONiC’s fabric CLI surface area (both the show fabric ... Click commands and the underlying fabricstat script) to expose additional fabric isolation/capacity telemetry and to add optional “non-zero only” filtering to reduce noise in command outputs.

Changes:

• Add ISOLATE_REASON display to fabric isolation output and introduce -nz/--nonzero filtering for isolation, port, queue, and rate views.
• Extend fabric capacity output to include isolated_links alongside operating/total links.
• Update unit tests and mock STATE_DB/COUNTERS_DB fixtures to cover the new columns and nonzero filtering behavior (primarily for fabricstat).

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/mock_tables/state_db.json	Adds isolated_links to fabric capacity fixture data.
tests/mock_tables/counters_db.json	Adjusts a counter fixture value to align with new nonzero filtering expectations.
tests/mock_tables/asic1/state_db.json	Adds isolation flags/reason fields and isolated_links for multi-ASIC fixture coverage.
tests/mock_tables/asic1/counters_db.json	Adjusts FEC uncorrectable counters to support nonzero-delta filtering tests.
tests/mock_tables/asic0/state_db.json	Adds isolation flags/reason fields and isolated_links for asic0 fixture coverage.
tests/mock_tables/asic0/counters_db.json	Adjusts a counter fixture value to align with nonzero filtering expectations.
tests/fabricstat_test.py	Expands expected outputs and adds new test cases for fabricstat -nz modes.
show/fabric.py	Adds -nz/--nonzero options to show fabric commands and forwards to fabricstat.
scripts/fabricstat	Implements isolate reason display, isolated link count in capacity output, and nonzero filtering logic.

[vmittal-msft]

Thanks for the change — I went through the full diff. Overall it looks good and the logic is consistent. A few notes below: one bug-fix call-out, agreement with the two automated comments, and a few items those didn't cover.

Nice catch / bug fix

Moving the rxRate/txRate/rxData/txData/time/local_time initialization inside the per-port loop in FabricRate.rate_print fixes a real state-leak — ports missing rate fields previously inherited the prior port's values. Confirmed against the mock data (ports 1/3/5 have no OLD_*_RATE_AVG, and the old expected output was asserting those leaked values). Good fix.

On the two automated review comments — both valid

1. Missing click-layer tests for -nz. fabricstat_test.py exercises the script directly via subprocess, so the show fabric ... click forwarding (if nonzero: cmd += ['-nz']) for port/queue/rate/isolation is untested. tests/show_test.py::TestShowFabric already covers port/queue — worth adding -nz variants there.
2. N/A treated as non-zero. _fabric_port_row_has_nonzero_counter returns True for None (N/A), so N/A rows are always shown while the help text says "non-zero". Reasonable behavior, but the help could call it out.

A few additional items

• Line length: scripts/fabricstat ~line 394 (the wrapped table_cnt.append(...)) is 123 chars, over the repo's --max-line-length=120. It was added in the first commit, so the diff-based pre-commit hook won't flag it, but a full flake8 run will.
• Inconsistent empty-table behavior under -nz: rate_print returns and prints nothing when the filtered body is empty, while isolation/port/queue still print a header with no rows. Minor, but worth making uniform.
• -nz semantics differ by view: for the full port/queue views the delta is computed vs. saved/cleared stats, so without a prior -C the delta equals the absolute counter and nearly every active link shows — it's most meaningful after a clear there (isolation/rate reflect current state regardless). Might be worth a doc note.
• -nz is silently ignored with -c (capacity) and reachability — fine, just undocumented.

I also confirmed -nz does not collide with -n/namespace in either argparse or click (exact registered option strings win), so no problem there.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[saksarav-nokia]

/azpw run

[mssonicbld]

⚠️ Notice: /azpw run only runs failed jobs now. If you want to trigger a whole pipline run, please rebase your branch or close and reopen the PR.
💡 Tip: You can also use /azpw retry to retry failed jobs directly.

Retrying failed(or canceled) jobs...

[mssonicbld]

Retrying failed(or canceled) stages in build 1152028:

✅Stage Build:

• Job Python3: retried.

[saksarav-nokia]

Thanks for the change — I went through the full diff. Overall it looks good and the logic is consistent. A few notes below: one bug-fix call-out, agreement with the two automated comments, and a few items those didn't cover.

Nice catch / bug fix

Moving the rxRate/txRate/rxData/txData/time/local_time initialization inside the per-port loop in FabricRate.rate_print fixes a real state-leak — ports missing rate fields previously inherited the prior port's values. Confirmed against the mock data (ports 1/3/5 have no OLD_*_RATE_AVG, and the old expected output was asserting those leaked values). Good fix.

On the two automated review comments — both valid

1. Missing click-layer tests for -nz. fabricstat_test.py exercises the script directly via subprocess, so the show fabric ... click forwarding (if nonzero: cmd += ['-nz']) for port/queue/rate/isolation is untested. tests/show_test.py::TestShowFabric already covers port/queue — worth adding -nz variants there.
2. N/A treated as non-zero. _fabric_port_row_has_nonzero_counter returns True for None (N/A), so N/A rows are always shown while the help text says "non-zero". Reasonable behavior, but the help could call it out.

A few additional items

• Line length: scripts/fabricstat ~line 394 (the wrapped table_cnt.append(...)) is 123 chars, over the repo's --max-line-length=120. It was added in the first commit, so the diff-based pre-commit hook won't flag it, but a full flake8 run will.
• Inconsistent empty-table behavior under -nz: rate_print returns and prints nothing when the filtered body is empty, while isolation/port/queue still print a header with no rows. Minor, but worth making uniform.
• -nz semantics differ by view: for the full port/queue views the delta is computed vs. saved/cleared stats, so without a prior -C the delta equals the absolute counter and nearly every active link shows — it's most meaningful after a clear there (isolation/rate reflect current state regardless). Might be worth a doc note.
• -nz is silently ignored with -c (capacity) and reachability — fine, just undocumented.

I also confirmed -nz does not collide with -n/namespace in either argparse or click (exact registered option strings win), so no problem there.

Addressed these review comments

[saksarav-nokia]

/azpw run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 1 pipeline(s).

[abdosi]

@kenneth-arista and @vmittal-msft to sign-off

[vmittal-msft]

Thanks for addressing the earlier round — the click-layer -nz tests (port/queue/rate/isolation), the empty-table early-returns, the semicolon (E703) cleanup, and the N/A note in help text all look good.

A few small nits remain. None fail CI (the pre-commit hook only lints the last commit's HEAD^..HEAD diff), but a full flake8 --max-line-length=120 scripts/fabricstat flags two of them:

1. scripts/fabricstat:625 — line too long (128 > 120, E501). The new capacity_header = [...] list is 128 chars. This is the same >120 line-length class as the earlier table_cnt.append note, just reintroduced on the capacity header. Suggest wrapping the list elements across lines.

2. scripts/fabricstat:205 — block comment should start with # (E265). #If the stats was not cleared before... — add a space after #. (This one was added in the "Addressed review comments" commit; the -U0 diff-based flake8 happens to miss it.)

3. scripts/fabricstat:564 — help-text typos. ' (rows with N/A are always shown);. Nout used with capacity and reachability ' has "Nout" (→ "Not") and a stray ;.. Minor, but user-visible in fabricstat -h.

4. Queue empty-table under -nz is still inconsistent. FabricPortStat/FabricIsolation/FabricRate now early-return and print nothing when the filtered result is empty, but FabricQueueStat.cnstat_print still prints a bare header with no rows. Worth adding the same if nonzero_only and len(table) == 0: return before the tabulate(...) for uniformity.

All cosmetic/polish — no functional concerns. Happy to approve once these are cleaned up (or if you'd prefer to defer #4 as a follow-up, that's fine).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[mssonicbld]

This PR has backport request label(s) for branch(es): msft-202601, but is missing required test information. Please make sure you tick the tested branch(es) in the Tested branch section and provide test evidence (e.g., 202601: <test result>) in the Test result section as well in your PR description.

_{---Powered by SONiC BuildBot}

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[saksarav-nokia]

@vmittal-msft @kenneth-arista @jfeng-arista , Please review this and sign off

[vmittal-msft]

Re-review (head 5b654f9) — following up on the Jul 15 round.

Thanks for the updates. 3 of the 4 items are resolved:

• ✅ capacity_header line length (E501) — wrapped across lines; a full flake8 --max-line-length=120 scripts/fabricstat now flags no PR-added line.
• ✅ # If the stats was not cleared... block comment (E265) — fixed (space after #).
• ✅ Queue empty-table under -nz — FabricQueueStat.cnstat_print now early-returns when the filtered table is empty, consistent with Port/Isolation/Rate.

One item is only partially addressed:

• ⚠️ -nz help text (scripts/fabricstat ~L570): Nout→Not is fixed, but the stray ;. I flagged earlier is still present (...always shown);. Not used...), plus a double space where the two string fragments join. Both are user-visible in fabricstat -h. Concrete suggestion inline.

Optional (non-blocking): in FabricRate.rate_print, rxData/txData/local_time are computed but unused now (body.append uses only rxRate/txRate) — leftover dead reads from the state-leak fix; drop if convenient.

No new lint or functional regressions in the latest commits. Happy to approve once the help string is cleaned up (or if you'd prefer to defer the cosmetic bits, that's fine too).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[mssonicbld]

The change is not in msft-202601 yet. @saksarav-nokia, please manually create the cherry pick PR for branch msft-202601.
You can ping the release branch owner(github account: rlhui) to approve your cherry pick PR.
If this change is already in msft-202601, please comment "already in msft-202601". Thanks!

_{---Powered by SONiC BuildBot}

[saksarav-nokia]

PR for msft-202601
Azure/sonic-utilities.msft#420