MWPW-185892: Show referring collections and projects in fragment details - #1199
Open
Axelcureno wants to merge 21 commits into
Open
MWPW-185892: Show referring collections and projects in fragment details#1199Axelcureno wants to merge 21 commits into
Axelcureno wants to merge 21 commits into
Conversation
- add getReferencedByFragmentId (GET by-id referencedBy) to the AEM client; it returns link-ready parents with id and model, plus cursor pagination - new references-repository: filters out pzn/promo/self-locale/cross-surface refs, groups collections by locale, buckets bulk-publish projects separately - render a "Referenced by" section in the fragment editor: collections link to the card editor, projects link to the bulk-publish editor, with locale counts and status chips - monotonic load-token plus abort controller guard against rapid fragment switches
…oxy)
- usage-repository: pure buildUsageQuery + parseUsageResponse for the Grafana
ClickHouse ds/query (org 750, trafficpeak read-only), unit-tested with mocked
response fixtures, no token or network needed
- fetchFragmentUsage targets a future IO Runtime proxy action and degrades to
{ available: false } on any failure (no proxy yet, auth, CORS, network)
- gated "Usage" section in the fragment editor: renders only when usage data is
available, so it stays invisible until the proxy exists and never affects the
shipped reference list
- the glsa_ service token must live only in the IO action inputs, never the
browser; direct browser to Grafana is both a secret leak and CORS-blocked
- new io/studio fragment-usage action: validates the caller IMS token, builds the
ClickHouse query server side (client cannot inject SQL), and forwards to Grafana
with the glsa_ service token read from the GRAFANA_TRAFFICPEAK_TOKEN deploy input
- the token lives only in the action input, never in source, logs, or the response;
action test asserts it reaches the Grafana auth header but never leaks in the body
- studio usage-repository now posts { fragmentId } to the action and parses the
Grafana response; still degrades to { available: false } until the action is
deployed, so the shipped reference list is unaffected
- default Grafana host set to adobe-grafana.trafficpeak.live (old grafana-us retired)
…tion - rename the "Referenced by" block to "References" and move it below grouped variations in the editor - bucket references by model type in display order: Collections, Cards, Promo Projects, Bulk Publish Projects, Localization Projects, Other - each type is its own collapsible row with a count, collapsed by default - promo and localization project rows deep-link to their editors; add buildPromoProjectDeepLink and buildTranslationProjectDeepLink
Match the deploy .env var name (GRAFANA_SERVICE_TOKEN / GRAFANA_SERVICE_URL) so the fragment-usage action's inputs resolve. Deployed and verified: both inputs inject from .env, endpoint returns 401 without a bearer.
…ndly) Usage hits the shared Grafana/Hydrolix cluster, so stop auto-loading it on every fragment open. It is now a collapsed "Usage" section below References that fetches only when the author expands it, and caches the result per fragment id for the session so re-expanding never re-queries. Cuts query volume from one-per-fragment- browsed to one-per-fragment-actually-inspected.
Constrain the References container to a fixed width and reduce vertical padding so it sits cleanly below grouped variations.
The epic's assumed schema (fragment_id/locale/api_key columns, datasource fdyta6qpga2o0d) does not exist on the migrated adobe-grafana instance. Derived the real shape from the "M@S - web_commerce_artifact" dashboard: - datasource ffmjsr3rpsrnkc, table akamai.logs, time column reqTimeSec - fragment id and api_key live in the query string (queryStr), not columns - match a fragment by reqPath/queryStr LIKE, group by api_key + country - default window tightened 30d -> 7d, plus hdx exec-time cap and an attribution comment, since a per-fragment LIKE scan on raw CDN logs is not cheap Deploy intentionally held pending ops sign-off on query scale. The reqTimeSec filter form and api_key extraction were derived from the dashboard, not executed (cluster probes are gated), so validate on first live run.
FluffyJaws (M@S monitoring wiki + MWPW-185891 + Slack) corrected the query the dashboard guess produced: - endpoint is reqPath = '/mas/io/fragment' (exact), not the web_commerce_artifact WCS path - extract id/locale/api_key from the query string with extractURLParameter, and match the fragment by exact id, not a substring LIKE (avoids false positives) - group by api_key, locale, country (locale is a real dimension); referer-page URLs stay out of scope per the story - deploy still held pending ops sign-off on scale Design note for the feature owner: the story's intended shape is a materialized hourly/daily/weekly rollup behind an IO action refreshed ~hourly, not a raw-log scan per panel open. This prototype does on-demand + per-session cache, which is fine for v1 but the rollup is the scalable target.
Match mas-related-variations: each type/usage collapsible is a standalone gray-50 rounded card with the same toggle styling, instead of one bordered box with thin inner rows. Also fix the usage label to "last 7 days" (matches the action's default window).
Add grafana_service_token to the shared-service-secrets fetch from merch-at-scale/aio-studio/common. The vault-secrets action uppercases it to GRAFANA_SERVICE_TOKEN, which the fragment-usage action reads. Keeps the token in Vault only, out of .env and the repo.
Local Studio can point the IO base at any personal namespace via ?io.ns= (e.g. merchatscale-axel when a dev deploy lands there instead of masstudio-axel). Gated to localhost so prod Studio can never be redirected to another runtime host.
…ct ternary - parseUsageResponse: rename ai/lo/ci/ni/f to descriptive index names. - toggleReferenceType: if/else instead of a ternary used for its side effect. No behavior change.
The referencedBy endpoint returns model.path as the plain path (the *_MODEL_ID constants are the base64 of that path, used by the fragment-model API). Matching promo / bulk-publish / localization projects by model.id never hit, so they all fell into the non-linkable "Other" bucket. Match by model.path, like collections and cards, so those rows classify correctly and get their deep links. Also reword the empty usage state to "No usage found for this fragment (last 7 days)"; update fixtures to the real model.path shape.
|
Hello, I'm the AEM Code Sync Bot and I will run some actions to deploy your branch.
|
Resolve studio/src/utils.js import conflict by keeping both PAGE_NAMES (this branch) and PZN_FOLDER (main); both are used in the file.
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #1199 +/- ##
==========================================
+ Coverage 90.71% 90.83% +0.11%
==========================================
Files 318 321 +3
Lines 101712 102513 +801
==========================================
+ Hits 92271 93118 +847
+ Misses 9441 9395 -46
... and 9 files with indirect coverage changes Continue to review full report in Codecov by Harness.
🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a References section to the Studio fragment details panel: the collections, promo / bulk-publish / localization projects, and cards that reference the open fragment, grouped by type in collapsible sections, each row a link. Uses the AEM GET-by-id
referencedByendpoint (returnsid+model, so rows are link-ready in one call rather than 1+N fetches). References classify bymodel.path, consistent across all types. Epic MWPW-186018, layer 4B.Also prototypes (epic 4A / MWPW-185891) a collapsed Usage (last 7 days) section: consumer / locale / country request counts from Akamai CDN logs via a new
fragment-usageIO proxy that keeps the Grafana service token server side. Marked PROTOTYPE. It degrades to hidden when the proxy or token is absent, so it never affects the shipped References list.Resolves https://jira.corp.adobe.com/browse/MWPW-185892
QA Checklist: https://wiki.corp.adobe.com/display/adobedotcom/M@S+Engineering+QA+Use+Cases
Please do the steps below before submitting your PR for a code review or QA
Test URLs:
Reviewer notes:
fragment-usageIO action. The Grafana service token is now sourced from Vault (aio-studio/common) in CI. Not exercised by the standard studio URL until deployed to the masstudio namespace; validated live against a personal namespace.references-repository.jsare exported and unit-tested but not yet wired into production; kept as covered helpers.