docs: add Generic HTTP Sink connector page and update connectors table

[mlevkov]

Context

Apache Iggy 0.8.0 (release post) ships the new Generic HTTP Sink connector (apache/iggy#2925). The connector's documentation currently lives only in the upstream core/connectors/sinks/http_sink/README.md and is not surfaced on the docs site.

The site's Available Connectors table does not list it, and there is no per-connector page under /docs/connectors/sinks/ analogous to the existing Elasticsearch, Postgres, MongoDB, Quickwit, Iceberg, or Stdout pages.

Proposed Changes

A scoped 3-file PR against main:

1. Add content/docs/connectors/sinks/http.mdx — new page sourced from the upstream README, reformatted to match the style of sibling sink pages.
2. Edit content/docs/connectors/sinks/meta.json — insert "http" in the pages array (controls sidebar order).
3. Edit content/docs/connectors/introduction.mdx — add the connector to the "Available Connectors" Sink row.

Naming & Style Decisions

Surface	Value	Rationale
File / URL slug	http.mdx → /docs/connectors/sinks/http	Matches sibling convention (no _sink suffix).
Page title (frontmatter)	HTTP Sink	Matches sibling pages: Elasticsearch Sink, Postgres Sink, etc.
Available Connectors table cell	Generic HTTP	"HTTP" alone is ambiguous — Elasticsearch and Quickwit sinks also use HTTP transport. "Generic" disambiguates as bring-your-own-endpoint, not target-specific. Matches the upstream commit and 0.8.0 release notes.
Sidebar / table position	Just before stdout	Existing order is roughly databases → search/lake → utility. Generic HTTP is utility-flavored, alongside stdout.

Resulting meta.json:

"pages": ["sink", "elasticsearch", "postgres", "mongodb", "quickwit", "iceberg", "http", "stdout"]

Resulting table row:

| Sink | PostgreSQL, MongoDB, Elasticsearch, Quickwit, Apache Iceberg, Generic HTTP, Stdout |

Content Scope (Proposed)

The upstream README is ~810 lines, well above any existing connector page on the site (range: 65–190 lines). Proposed scope for http.mdx, targeting ~200 lines:

Include (essential surface):

• Intro paragraph
• Full Configuration options table
• Batch Modes (individual, ndjson, json_array, raw) — unique to this connector
• Metadata Envelope (output shape)
• Authentication patterns (Bearer / API key / Basic)
• Retry strategy summary + Delivery Semantics caveat (at-most-once)
• 2–3 representative use-case configs (webhook, bulk ingestion, IoT/raw)
• Known Limitations

Summarize with a link back to the upstream README:

• Deployment Patterns (multi-instance, fan-out, Docker recipes)
• Performance Considerations (connection pooling, batch-mode throughput math)

Omit:

• "Try It" hands-on tutorial — the runtime page already covers how to run a connector
• Testing instructions — repo-developer concern

Source-of-Truth Strategy

Upstream README.md remains canonical. The site page is a curated reference subset that links back to the upstream README for the sections we summarize. We don't mirror line-for-line because the README also serves as a contributor doc with internal source links (e.g. runtime/src/sink.rs) that don't belong on the public site. Future connector changes should update both the upstream README and this page in lockstep.

References

• Upstream PR: apache/iggy#2925
• Release post: Apache Iggy 0.8.0
• Upstream README: core/connectors/sinks/http_sink/README.md