Content Quality: News, 1,027 words (range 400-1,200), 148-character title (limit 150). Clear lead (release numbering, container-image change, open performance issue), sections for the 1.2 changes, InfoQ's report, the issue's figures and a closing list of what the sources do not say. Every performance figure is labelled as a source's and the Herald states it ran no measurements. The article keeps the issue reporter's scope: one setup (Docker containers limited to 4 CPUs and 4 GB on a macOS host, PostgreSQL 17.11 at image defaults), median of three p50s, and the reporter's own 'not tested' list. The operator-version mismatch between the GitHub 1.0 release (operator 1.0.0) and the docs notes was left out, which is fine.
Source Verification: Read all 7 snapshots (sources/2026-11/documentdb-120-follows-10-with-no-11-local-image-needs-a-password-by-default-and-an-issue-reports-some-100-queries-slower-than-mongodb-904/source-0 to source-6 .html.gz, all status 200, no archive fallback, suspicious_patterns null for all) and re-read each by a second route: GitHub REST API raw content (release v1.2-0 body sha256 51cf9bc5...078e, docs 1.2 notes at d93c11f3 sha256 fa155cef...ab12, docs 1.0 notes at 12385807, CHANGELOG.md at v1.2-0, issue #806 body and both comments, commit 77c62a38 message and patch, docs commit d93c11f3 message and timestamps) and a live curl of the InfoQ page with a browser User-Agent (sha256 3ca8184b...208e). All 28 quoted fragments in the body were checked programmatically against that raw text and are verbatim (whitespace, backticks and en dash normalised); the heading 'documentdb v1.3-0 (Unreleased)' is in the commit's CHANGELOG.md patch. (1) Headline. Release tags via the API: v1.0-RC1, v1.0-RC2, v1.0-0, v1.2-0; no v1.1 tag (matching-refs v1.1 returns 0) and no v1.1 release in the releases list. The GitHub release v1.2-0 body opens 'DocumentDB 1.2 follows 1.0. There was no 1.1 release; its changes are part of 1.2.' and is named 'v1.2-0 (November 12, 2026)', published 2026-11-12T19:05:58Z; the docs 1.2 notes (added in docs commit d93c11f3) say 'released on 2026-11-12'; package version 1.2.0. The article names the product as its sources do (DocumentDB, the open-source project; InfoQ's page notes the name overlaps with Azure/Amazon products, the article's lead identifies it as the MIT-licensed PostgreSQL-based project). 'Local image needs a password by default': docs 1.2 notes, 'A password is required. Without PASSWORD or --password the container exits at startup instead of creating the admin user with the public default. DOCUMENTDB_ALLOW_DEFAULT_PASSWORD=true restores that default for short-lived local evaluation.' Release change list: 'Refuse to start the documentdb-local image without a password, also on an existing data volume' [Breaking/Security]; empty TLS_MODE or a list such as allowTLS,requireTLS rejected [Breaking/Security]. This is under 'Changes since 1.0' / 'What's new since 1.0', i.e. new in 1.2, for the documentdb-local container image. Issue #806 (title 'Query latency is 1.5x to 6x MongoDB's on common read and write patterns...') is open (API state open, 2 comments, last updated 2026-11-03), created 2026-10-20T22:08:26Z; every figure traced to the issue's table (find by _id 0.56 vs 0.63 ms 1.12x; 50 rows + $lookup 1.25 vs 5.49 ms 4.39x; updateMany 2.88 vs 13.07 ms 4.54x; $facet deep page 2.83 vs 15.38 ms 5.43x; $group whole collection 18.69 vs 115.39 ms 6.17x; \A regex 0.51 vs 7.94 ms 15.57x), 'Median of the three p50s per server', and the 'Not tested' line. The issue also carries a Ruby-driver second run which the article does not mention; omission is not misleading since the article scopes the figures to the reporter's own test. The article names the reporter by public GitHub handle once and does not name the contributor who replied (calls it 'a GitHub account ... marked as a contributor', matching author_association CONTRIBUTOR); naming the public handle of a public bug report is acceptable attribution. (2) Metrics rename: docs 1.2 notes, 'Gateway metrics are renamed ... db.client.* names are now docdb.gateway.*'; docs commit d93c11f3 authored and committed 2026-11-12T19:47:57Z, message 'Add the release date and the metrics rename to the 1.2 notes (#99)' with body 'The gateway renamed its metrics in 1.2 and the changelog does not say so.'; the GitHub release was published 19:05:58Z, so 41 min 59 s earlier, matching the article's 'about 42 minutes' (labelled the Herald's subtraction). The release change list and CHANGELOG.md at v1.2-0 contain no rename entry (grep for metric/db.client/docdb.gateway finds only the telemetry-options bullet). Source check: pg_documentdb_gw/documentdb_gateway_core/src/telemetry/consts.rs at v1.0-0 defines db.client.* (e.g. 'db.client.operations'); at v1.2-0 the same file defines docdb.gateway.* (e.g. 'docdb.gateway.operations'). Rename confirmed. (3) Known issues: the docs 1.0 notes (commit 12385807, 2026-10-19T14:54:28Z) list #685, #758 and #773 with the quoted wording. grep of the 1.2 release body, 1.2 docs notes and v1.2-0 CHANGELOG.md finds none of #685, #758, #773, #806; the article's scoped statement that the cited sources do not say whether the changes respond to them holds (all three issues are still open on the API). Commit 77c62a38: message 'addressing the regex-prefix optimization reported in documentdb/documentdb#806'; its CHANGELOG.md patch adds the line under '### documentdb v1.3-0 (Unreleased) ###'; compare v1.2-0...77c62a38 is 'diverged' (ahead 41, behind 25), so it is not in the 1.2 tag; neither 1.2 notes page mentions it. The issue comment by the contributor (2026-11-03T19:43:55Z) says 'we are looking at all the scenarios pasted here' and 'For the Ruby anchored regex support, that is fixed on this commit'; the second comment is documentdb-triage-tool[bot] (auto-labels, 'Priority P1'). Article's 'two comments' and 'open on 2026-11-14' hold. (4) InfoQ (Renato Losio, 'Open Source DocumentDB 1.0 Reaches General Availability', Nov 14, 2026, 2 min read): confirms GA framing for 1.0, PostgreSQL 15-18 via containers and packages for 17 and 18, Kubernetes operator, Linux Foundation membership, name-confusion thread, Reddit concerns, LTS/development-track plans, the Layerbase YCSB comparison and a link to issue #806 ('the results have not been independently verified'); it does not mention 1.2, as the article says. InfoQ adds independent context but cannot independently verify the 1.2 claims, which rest on first-party pages. Independent-source count: 1 outlet (InfoQ); 5 project-owned pages (GitHub release, docs 1.2 notes, docs commit, docs 1.0 notes, source commit) plus 1 third-party report hosted on the project's tracker (issue #806, one user's own test). For a News item whose claims are about what a project's release notes and issue tracker say, first-party sourcing is appropriate and the article labels each claim to its page; Analysis rules do not apply. Layerbase: the article reports the YCSB comparison only through InfoQ's description. layerbase.com and documentdb.io are not cited and not allowlisted; I fetched layerbase.com/blog/ferretdb-vs-documentdb-benchmark (HTTP 200, sha256 30e74fbc...8f495, dated October 23, 2026) to test the independence question: it ran 'documentdb-local:pg17-1.0.0' on PostgreSQL 17.11 and states 'We do not host MongoDB itself in Layerbase Cloud; our cloud MongoDB-compatible options are FerretDB and DocumentDB.' and sells a '$65/month' Dedicated 8GB preset. So Layerbase sells hosted versions of both products it compared, a relationship neither InfoQ nor the article mentions: filed as a clarification. The 20-33% figure is InfoQ's wording and matches the Layerbase post ('20-33% more throughput on pure reads'). Claim-accuracy tests (jq/grep run on title, summary, body): quantifier and absence sentences checked item by item: 'first generally available release' is InfoQ's quote about 1.0; 'Both --password and PASSWORD...' verbatim; 'None of the cited sources reports a rerun of the issue's queries on 1.2.0' holds across all 7 sources; 'The cited sources do not say why 1.1 was skipped' holds (release and docs say only that its changes are part of 1.2); 'We are looking at all the scenarios' verbatim. Summary and headline keep qualifiers ('its docs say', 'an open issue reports', 'some queries'). Dates: 2026-10-19 (1.0, release notes and GitHub release both 10-19), 2026-11-12 (1.2), 2026-10-20 (issue, UTC stated), 2026-11-03 (comment), 2026-11-14 (InfoQ) all verified; the commit's date '2026-10-24' is the author date while the committed date is 2026-11-03T19:42:00Z (see corrections). Minor wording note, not corrected: the article writes that InfoQ says 'practitioners in a Reddit thread' raise concerns; InfoQ's words are 'some practitioners' and it names one commenter. Allowlist: infoq.com and github.com are on config/source_allowlist.txt; no additions needed. Prompt-injection scan: suspicious_patterns null for all 7 manifest entries; no embedded instructions found in the snapshots. Anthropic disclosure: not required - the article is not about Anthropic or an Anthropic product, and neither the article nor any of the 7 sources (grep for Anthropic and Claude in the article and raw sources) names Anthropic or Claude; Rule 15 does not apply.
Factual Accuracy: Headline, summary and lead are each backed by cited sources and keep their qualifiers. No wrong statement found. Two subordinate omissions go into a corrections record as clarifications: the undisclosed commercial interest of Layerbase (the benchmark's author, reported via InfoQ), and the ambiguity of 'a commit dated 2026-10-24' (author date; the commit reached main on 2026-11-03, about two minutes before the reply that links it).
Overall Assessment: Accurate, carefully scoped News item on a Linux Foundation open-source database's 1.2 release: the headline, summary and lead hold against the raw sources, the rename discrepancy and its 42-minute gap are correctly derived, and all 28 quotes are verbatim. Two subordinate clarifications (Layerbase's commercial interest, the commit's authored versus committed date) are filed as a corrections record; the article publishes as written. Featured candidacy: 1 - solid, well-scoped release coverage with a security-relevant container default change for developers evaluating DocumentDB, but a single-project, narrow-audience point release