Refactor: right align metrics email by MarceloRobert · Pull Request #1943 · kernelci/dashboard

MarceloRobert · 2026-06-18T12:55:05Z

Readjusts the email formatting so that the larger number sections have a right-aligned text.

Changes

Turned left-aligned text in metrics email's sections "COVERAGE" and "TEST LAB ACTIVITY" into right-aligned text;
Adjusted spacing for some columns;
Changed double space before change percentage into a single space (that's 123 (+99%) into 123 (+99%));
Removed percentage sign if the metric diff is 0. This was done because many labs don't have tests, so the (+0%) can clutter the change column.
Used dynamic spacing on Coverage and Test Labs Activity in order to avoid large empty spaces on all-small numbers and also avoid lack of space for larger-than-usual numbers

How to test

Have data in the database, I used /backend/tests_submissions/submissions.zip for testing
Run poetry run python3 manage.py notifications --action metrics_summary and compare with the older version

Visual Reference

Example of new format:

KernelCI Metrics Summary
========================
Period: 2026-06-13 17:33 UTC to 2026-06-20 17:33 UTC


  COVERAGE
  --------
                   This week  Last week   Change
                   ─────────  ─────────  ───────
  Trees monitored          4          0       +4
  Checkouts               32          0      +32
  Builds                 103          0     +103
  Tests               15,720          0  +15,720


  BUILD REGRESSIONS
  -----------------
  A "regression" is defined as a reported problem affecting 0 or multiple builds.

  Origin    Regressions       Affected           Affected builds by top issues
            (known + new)     Builds (total)     #1      #2      #3
  ─────────────────────────────────────────────────────────────────────────
  maestro   0   + 1   = 1     1                  1       
  ─────────────────────────────────────────────────────────────────────────
  Total     0   + 1   = 1     1


  TOP REGRESSIONS PER ORIGIN
  --------------------------
  maestro:
    #1 (1 occurrences) -  in vmlinux (Makefile:1220) [logspec:kbuild,kbuild.other]
      Dashboard: https://d.kernelci.org/issue/maestro:5dd754a8374e1b44291f37f6a3510603dcbc6348?iv=1
  

  TEST LABS ACTIVITY
  ------------------
  13 labs reported results this week (13 more than last week).

  Labs marked with an asterisk (*) are new.

  Lab                    Covered builds   Boots    Tests   Change (tests)
  ───────────────────────────────────────────────────────────────────────
  lava-baylibre *                     1       1        0                0
  lava-broonie *                     41       8    3,467           +3,467
  lava-cip *                          1       1        0                0
  lava-clabbe *                       3       3        0                0
  lava-collabora *                   82      29    1,805           +1,805
  lava-foundriesio *                  2       2        0                0
  lava-kci-qualcomm *                 1       0        1               +1
  lava-pengutronix *                  3       3        0                0
  linaro *                          174     229   10,119          +10,119
  maestro *                           1       0        2               +2
  microsoft *                         1       0        3               +3
  opentest-ti *                       2       2        0                0
  redhat *                            1       0       45              +45
  ───────────────────────────────────────────────────────────────────────
  Total                             313     278   15,442          +15,442

--
This is an experimental report format. Please send feedback in!
Talk to us at kernelci@lists.linux.dev

Made with love by the KernelCI team - https://kernelci.org

Closes #1938

alanpeixinho

You are the G.O.A.T. Nice work

alanpeixinho · 2026-06-22T13:50:35Z

    diff = cur - prev
    if diff == 0:
-        return "0  (0%)"
+        return "0"


Just a question: is there any reason we are not just following the show_percentage parameter here? to decide if we should show or not?

Also, can we return a tuple of int instead? Then the template can align them separately, and we further separate data processing from templating

I'm not sure tbh, but even still the lab changes are currently using percentages but they would be better off if the zeroes didn't have the percentage IMO

alanpeixinho · 2026-06-22T13:55:07Z

        percentage = round((diff / prev) * 100)
        percentage_sign = "+" if percentage > 0 else ""
-        return f"{signed_diff}  ({percentage_sign}{percentage}%)"
+        return f"{signed_diff} ({percentage_sign}{percentage}%)"


nit: perhaps we could use f"{diff:+} {percentage:+}" here

Sorry, I didn't get it. The difference between f"{signed_diff} ({percentage_sign}{percentage}%)" and f"{diff:+} {percentage:+}" would be from +100 (+200%) to +100 +200 for diff = 100 and prev = 50

I only meant to use '+' format modifier, but is not necessary.

mentonin

Thanks for tackling this. I like that you included dynamic field spacing too! I left some comments for improvements, but they are not blockers

mentonin · 2026-06-23T03:33:39Z

+  {{ "{:>{width}}".format("Change", width=change_space) }}
+  {#-#}
+  {{ "{:<{width}}".format("", width=label_space) -}}
+  {{ "{:>{width}}".format("─" * (current_week_space-static_spacing), width=current_week_space) -}}


should we make this horizontal rule consistent with the other tables'?

mentonin · 2026-06-23T03:35:47Z

+  {{ "{:>{width}}".format(deltas.n_tests, width=change_space) }}


  BUILD REGRESSIONS


BUILD REGRESSIONS should also be right aligned for consistency

I think it's not needed to change build regressions since the scale there is smaller, eg. https://groups.io/g/kernelci-results/message/61332

Maybe leave the auto-width out, but align the numbers right? the report you linked to still spans 3 orders of magnitude on Affected Builds

That works, done.

mentonin · 2026-06-23T03:40:09Z

+{%- set lab_builds_space = [ns.lab_builds_len, "Covered builds"|length] | max + static_spacing -%}
+{%- set lab_boots_space = [ns.lab_boots_len, "Boots"|length] | max + static_spacing -%}
+{%- set lab_tests_space = [ns.lab_tests_len, "Tests"|length] | max + static_spacing -%}


You can drop these if you initialize ns with the label lengths instead of 0

mentonin · 2026-06-23T03:42:58Z

+{%- set lab_change_len = deltas.labs.values() | map('length') | max -%}
+{%- set lab_change_space = [lab_change_len, "Change (tests)"|length] | max + static_spacing -%}
+
+{%- set separator_length = lab_names_space + lab_builds_space + lab_boots_space + lab_tests_space + lab_change_space - 2 -%}


nit: this is often called an horizontal rule (or rule, hrule, or hr) instead of the more generic "separator"

mentonin · 2026-06-23T03:45:57Z

+{#- Compute the text spacing in jinja instead of python to consider the `fmt` macro -#}
+{%- set static_spacing = 3 -%}
+{%- set lab_names_len = lab_maps.keys() | map('length') | max -%}
+{%- set lab_names_space = [lab_names_len, "Lab"|length ] | max + static_spacing + 2 -%} {#- +2 for the possible ` *` -#}


nit: group all *_space for a table together

Wdym by grouping? just remove the empty lines between the variables? I think it helps to discern which steps are happening and which order they'll appear

mentonin · 2026-06-23T03:53:35Z

    diff = cur - prev
    if diff == 0:
-        return "0  (0%)"
+        return "0"


Also, can we return a tuple of int instead? Then the template can align them separately, and we further separate data processing from templating

mentonin · 2026-06-23T03:56:51Z

+  {{ "{:>{width}}".format(fmt(0), width=lab_tests_space) -}}
+  {{ "{:>{width}}".format(deltas.labs.get(lab_key, ""), width=lab_change_space) }}
+{% endfor %}
+{{- "  " + "─" * separator_length}}


formatting - can (should) we add jinja template formatting to the project?

Unfortunately it's hard to get a good formatting here because if I just remove the " " it will be misaligned, if I add a space before {{"─" * separator_length}} it will contain the line break after {% endfor %}, and if I add the space there but add a -%} at the endfor to ignore the line break it will also ignore the space which will make it misaligned anyway.

I could of course keep it the way it was ({% endfor %} {{ "─" * separator_length}}) but IMO it looks worse having both commands in the same line

mentonin · 2026-06-24T21:00:00Z

LGTM

Readjusts the email formatting so that the larger number sections have a right-aligned text Part of kernelci#1938 Signed-off-by: Marcelo Robert Santos <4mrSantos@gmail.com>

Calculates the space required for the data and labels within jinja to avoid empty spaces or crammed values in the Coverage and Test Labs Activity sections It is possible to do it in Python but then we would need the fmt macro outside as well. A value such as 1000 might count as 4 chars in python but it is displayed as a 5-char 1,000 in jinja. Also, these spaces only make sense for the jinja template anyway, so I think it is correct to keep them within it. Signed-off-by: Marcelo Robert Santos <4mrSantos@gmail.com>

Closes kernelci#1938 Signed-off-by: Marcelo Robert Santos <4mrSantos@gmail.com>

MarceloRobert · 2026-06-25T12:12:41Z

@mentonin I refactored the build regressions section to right align there too, but I'm using static spacing for now

MarceloRobert marked this pull request as ready for review June 20, 2026 17:34

MarceloRobert changed the title ~~[WIP] Refactor: right align metrics email~~ Refactor: right align metrics email Jun 20, 2026

MarceloRobert force-pushed the refactor/right-align-metrics branch from a1270f2 to fb68cfd Compare June 20, 2026 17:59

alanpeixinho approved these changes Jun 22, 2026

View reviewed changes

mentonin reviewed Jun 23, 2026

View reviewed changes

MarceloRobert force-pushed the refactor/right-align-metrics branch from fb68cfd to 6ffa3e6 Compare June 24, 2026 12:22

mentonin approved these changes Jun 24, 2026

View reviewed changes

MarceloRobert added 3 commits June 25, 2026 09:11

refactor: right align metrics email

0b82828

Readjusts the email formatting so that the larger number sections have a right-aligned text Part of kernelci#1938 Signed-off-by: Marcelo Robert Santos <4mrSantos@gmail.com>

refactor: right align build regressions in metrics email

381f98a

Closes kernelci#1938 Signed-off-by: Marcelo Robert Santos <4mrSantos@gmail.com>

MarceloRobert force-pushed the refactor/right-align-metrics branch from 6ffa3e6 to 381f98a Compare June 25, 2026 12:11

		{{ "{:>{width}}".format(deltas.n_tests, width=change_space) }}


		BUILD REGRESSIONS

Uh oh!

Conversation

MarceloRobert commented Jun 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changes

How to test

Visual Reference

Uh oh!

alanpeixinho left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mentonin left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

MarceloRobert Jun 24, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mentonin commented Jun 24, 2026

Uh oh!

MarceloRobert commented Jun 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

MarceloRobert commented Jun 18, 2026 •

edited

Loading

MarceloRobert Jun 24, 2026 •

edited

Loading