feat(agentics): add daily performance-scan agentic workflow by sergio-sisternes-epam · Pull Request #2125 · microsoft/apm

sergio-sisternes-epam · 2026-07-10T14:40:40Z

TL;DR

Adds a daily agentic workflow (perf-scan) that scans src/apm_cli/ for
algorithmic performance anti-patterns and opens a GitHub Issue with findings
every morning at 01:00 UTC.

Stacked on #2124 -- do not merge until #2124 merges first.
Depends on .github/agents/performance-expert.agent.md and
.github/agents/algorithmic-patterns.agent.md introduced in that PR.

Problem (WHY)

PR #2124 landed four algorithmic perf fixes and a performance-expert agent
with a full algorithmic lens. Without a recurring automated scan, new regressions
will silently accumulate between manual audits.

Approach (WHAT)

A GitHub Agentic Workflow that:

Runs daily at 01:00 UTC (results ready before 6-7am UK/CET) and on workflow_dispatch.
Scans six anti-pattern categories in src/apm_cli/.
Always creates a GitHub Issue (findings or clean) so the run is always visible.

Implementation (HOW)

Files added

File	Purpose
`.github/workflows/perf-scan.md`	Agentic workflow: frontmatter + agent instructions
`.github/workflows/perf-scan.lock.yml`	Compiled YAML (generated by `gh aw compile v0.80.9`)
`.github/agents/workflows/perf-scan.md`	Companion reference: scope, output contract, finding format

Anti-patterns scanned (A-F)

ID	Pattern	Typical impact
A	Quadratic loop nesting	O(n) -> O(n^2) as package count grows
B	Linear scan in loop (`x in list`)	O(n) per iteration vs. O(1) with set/dict
C	Unconditional expensive ops on hot paths	Avoidable I/O per call
D	Redundant env-var / config parsing	Repeated syscalls per invocation
E	Heavy top-level imports in command modules	Startup latency regardless of sub-command
F	Sequential independent I/O	Wall-time parallelism opportunity

Schedule and output

Schedule: 0 1 * * * (01:00 UTC daily) + workflow_dispatch.
Issue titles: [perf-scan] YYYY-MM-DD -- performance opportunities found or -- no issues found
Labels: performance, automated. ASCII-only body.

Security

Issue creation via safe-outputs: create-issue only. No issues: write direct permission.

Validation

gh aw compile perf-scan
[+] .github/workflows/perf-scan.md (53.9 KB)
[!] Compiled 1 workflow(s): 0 error(s), 1 warning(s)

One expected warning: fixed 01:00 UTC cron (intentional -- daily fuzzy schedule fires at a random time).

How to test

Actions > Daily Performance Scanner > Run workflow (manual dispatch).
Confirm a new issue labelled performance + automated appears in the tracker.

…ss ref resolution - dependency_graph.py: Replace O(n) has_dependency() linear scan with O(1) set lookup via _repo_url_index maintained on add_node() - http_cache.py: Add tracked size fast-path to _enforce_size_cap() so the full directory scan is skipped when total cache is under the 100MB cap (the common case on every store() call) - marketplace/builder.py: Replace triple sequential iteration over git refs list (tags, full refnames, branches) with a single-pass dict index built once per resolution, reducing O(3n) to O(n) + O(1) lookups - github_host.py: Extract repeated os.environ.get + strip + lower + split chains into shared helpers (_get_ghes_host, _get_gitlab_single_host, _get_gitlab_hosts_list) to eliminate redundant string processing on every host classification call Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Cover the three main optimisation areas: - has_dependency() O(1) index vs O(n) scan scaling - HttpCache size-cap fast-path (linear scaling, no quadratic rescan) - github_host env-var helpers throughput Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

…ivation Add Big O analysis, missing-index detection, unconditional-expensive-op flagging, import-cost awareness, and scaling-guard recommendations to the performance-expert agent. Create references/algorithmic-patterns.md as a lazy-loaded catalogue of the six core anti-patterns. Broaden the panel activation rules to fire on utils/, marketplace/, compilation/ paths and on diffs that introduce loops over collections or repeated I/O calls. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Add perf-scan.md agentic workflow that runs daily at 01:00 UTC and scans src/apm_cli/ for six algorithmic performance anti-patterns: A. Quadratic or worse loop nesting B. Linear scan inside a loop (set/dict opportunity) C. Unconditional expensive operations on hot paths D. Redundant env-var / config parsing in sequence E. Heavy top-level imports on CLI command modules F. Sequential I/O with no data dependency (parallelism opportunity) Each run creates a GitHub Issue titled: [perf-scan] YYYY-MM-DD -- performance opportunities found [perf-scan] YYYY-MM-DD -- no issues found (when clean) Issues are labelled 'performance' and 'automated'. Output is ASCII-only per the repository encoding rule. The companion instructions file at .github/agents/workflows/perf-scan.md documents the agent prompt, scan scope, and output contract for reference. The .lock.yml was generated with gh aw compile (v0.80.9). Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Adds a new gh-aw agentic workflow (perf-scan) scheduled daily at 01:00 UTC to scan src/apm_cli/ for algorithmic performance anti-patterns and open a tracking issue per run, plus a companion reference doc.

Changes:

Introduces .github/workflows/perf-scan.md with the scan procedure and issue-format contract.
Adds the compiled workflow .github/workflows/perf-scan.lock.yml.
Adds .github/agents/workflows/perf-scan.md as a standalone reference for scope and output contract.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.

File	Description
.github/workflows/perf-scan.md	Defines the scheduled agent instructions + safe-output issue creation contract.
.github/workflows/perf-scan.lock.yml	Generated workflow YAML used by GitHub Actions to execute the agentic workflow.
.github/agents/workflows/perf-scan.md	Companion reference describing scope, output contract, and finding format.

+(The `[perf-scan] ` prefix is added automatically by the safe-output.)
+
+Create the issue using the `create-issue` safe-output tool with the body below.
+All text must be ASCII-only.


+# This file was automatically generated by gh-aw (v0.65.6). DO NOT EDIT.
+#
+# To update this file, edit the corresponding .md file and run:
+#   gh aw compile
+# Not all edits will cause changes to this file.
+#
+# For more information: https://github.github.com/gh-aw/introduction/overview/
+#
+# Scans src/apm_cli/ daily for algorithmic performance anti-patterns and opens a GitHub Issue with findings
+#
+# gh-aw-metadata: {"schema_version":"v3","frontmatter_hash":"9540b93dbcfcc84ddf0ca18c81b0a736840a9a6517bfc634c0258941a709831e","compiler_version":"v0.65.6","strict":true,"agent_id":"copilot"}
+


+grep -rn --include="*.py" -E "if .+ in \[|\.count\(|\.index\(" src/apm_cli/ | \
+  grep -v "test_\|#\|isinstance" | head -60


+  create-issue:
+    expires: 7d
+    title-prefix: "[perf-scan] "
+    labels: [performance, automated]


+    runs-on: ubuntu-slim
+    permissions:
+      contents: read
+      issues: write
+    timeout-minutes: 15


+Every run creates exactly one GitHub Issue:
+- Found findings: `[perf-scan] YYYY-MM-DD -- performance opportunities found`
+- No findings: `[perf-scan] YYYY-MM-DD -- no issues found`
+
+Labels applied: `performance`, `automated`.
+


Sergio Sisternes and others added 4 commits July 10, 2026 14:04

chore: re-sync lockfile and deployed files after agent expansion

c21128c

Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>

Copilot AI review requested due to automatic review settings July 10, 2026 14:40

sergio-sisternes-epam requested a review from danielmeppiel as a code owner July 10, 2026 14:40

sergio-sisternes-epam force-pushed the sergio-sisternes-epam-daily-perf-scan-workflow branch from 294959c to 28f2b29 Compare July 10, 2026 14:41

sergio-sisternes-epam changed the base branch from main to sergio-sisternes-epam-perf-audit-and-fixes July 10, 2026 14:41

Copilot started reviewing on behalf of sergio-sisternes-epam July 10, 2026 14:41 View session

Copilot AI reviewed Jul 10, 2026

View reviewed changes

sergio-sisternes-epam force-pushed the sergio-sisternes-epam-perf-audit-and-fixes branch from 029b93e to c5c3fd1 Compare July 12, 2026 05:33

Base automatically changed from sergio-sisternes-epam-perf-audit-and-fixes to main July 12, 2026 05:46

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat(agentics): add daily performance-scan agentic workflow#2125

feat(agentics): add daily performance-scan agentic workflow#2125
sergio-sisternes-epam wants to merge 5 commits into
mainfrom
sergio-sisternes-epam-daily-perf-scan-workflow

sergio-sisternes-epam commented Jul 10, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		grep -rn --include="*.py" -E "if .+ in \[\|\.count\(\|\.index\(" src/apm_cli/ \| \
		grep -v "test_\\|#\\|isinstance" \| head -60

Uh oh!

Conversation

sergio-sisternes-epam commented Jul 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

TL;DR

Problem (WHY)

Approach (WHAT)

Implementation (HOW)

Files added

Anti-patterns scanned (A-F)

Schedule and output

Security

Validation

How to test

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

sergio-sisternes-epam commented Jul 10, 2026 •

edited

Loading