[braxtonj]

Checkly exposes metrics on one Prometheus endpoint and sends check runs as OpenTelemetry spans to one OTLP endpoint. The pages for that are split between Integrations > Observability and Resolve > Traces > Exporting today, and Chronosphere is not in the nav. This PR puts them in one place.

What changes

• Integrations > Observability now holds everything that leaves Checkly: an Overview with a decision table, a Metrics subgroup with one Prometheus page, and a Traces subgroup with a generic OTLP page, Coralogix and thin vendor pages (Chronosphere, ClickStack, Cribl Stream, Dash0, Datadog, Grafana Tempo, Honeycomb, New Relic).
• Resolve > Traces > Exporting becomes one router page that links the Integrations pages. No page path appears in two nav groups.
• Prometheus is one page. The v2 content is the page, with labels, windows and edge cases documented from the current implementation. The v1 endpoint is a clearly labeled legacy section at the end with a deprecation warning. /integrations/observability/prometheus-v2 redirects to /integrations/observability/prometheus and the #heartbeat-metrics anchor survives.
• The OTLP page is new and carries the span schema, the trace headers, the tags rules, a tested Collector receiver recipe, transport facts and troubleshooting. Vendor pages only carry the values each vendor needs and link back for everything else.
• Coralogix is corrected in place: key location, endpoint derivation, the attribute list, the custom trace header behavior, and current RUM key names.
• Resolve > Traces pages point at the new locations; how-it-works keeps its diagram and gets the four-type coverage wording. A vendor-neutral Mermaid version of the diagram sits on the page inside an MDX comment, ready to swap in.

Moves and redirects

Old	New
resolve/traces/export/grafana-tempo	integrations/observability/traces/grafana-tempo
resolve/traces/export/honeycomb	integrations/observability/traces/honeycomb
resolve/traces/export/new-relic	integrations/observability/traces/new-relic
resolve/traces/export/dash0	integrations/observability/traces/dash0
resolve/traces/export/to-chronosphere	integrations/observability/traces/chronosphere
integrations/observability/prometheus-v2	merged into integrations/observability/prometheus

All six have docs.json redirects, so existing links, including the ones in the web app, keep working.

Verification

• Schemas, labels and attributes were taken from the current implementation and checked live, not copied from the old pages.
• Live verification on 2026-10-30: spans captured for every check type from public and private locations through a bearer-authenticated OpenTelemetry Collector, the receiver recipe on the OTLP page. The scrape config, every PromQL example, the Collector prometheus receiver block and the legacy v1 config were run in Prometheus 3.0.0 and otelcol-contrib 0.160.0. Both Prometheus endpoints were scraped and compared with the documented metric families (35 of 35 v2, 6 of 6 v1).
• Vendor pages were validated against vendor documentation. Anything that could not be confirmed was left off.
• Screenshots: fresh Checkly-side captures of the Prometheus card, the Export Traces settings and the Coralogix form, values masked. The Export Traces capture is reused on the Dash0 and Grafana Tempo pages. The existing vendor-side screenshots on those pages and on Coralogix sit under the rewritten steps, and the Prometheus page keeps its two earlier setup captures. The token visible in dash0-token.png is masked.

Deliberately dropped

• Three old Checkly screenshots: dash0-checkly.png and export-traces-to-grafana-config.png (the old Export Traces form, replaced by the fresh capture) and prometheus_step2.png (the v1 activation card; v1 setup is now one sentence under Legacy v1 endpoint). The files stay on disk.
• Grafana's "check that Tempo is installed" step and the third-party base64 website (replaced by a shell one-liner).
• New Relic's "Traces feature enabled" step (not in New Relic's documentation).
• The v2 page's Info banner and the v1 Warning (replaced by one deprecation Warning), the v1 "inactive after six hours" sentence, the v1 "30 to 60 s interval" advice, and the enable_compression tip (the endpoint gzips when the scraper asks, which Prometheus does by default).
• A link to the private location agent metrics page: docs(private-locations): agent Prometheus metrics and single server port #755 is still open. Once it lands, add one sentence on the Prometheus page and point its Prometheus link at /integrations/observability/prometheus.
• The deprecated agentic check type from the check_type enumeration.

Follow-up

• Redraw checkly_otel_diagram.png on how-it-works: it hardcodes tracestate: checkly=true and lists Baselime among the vendors. The Mermaid version on the page is the starting point.

Open PRs that touch the same files

• docs: sentence case for headings and titles #774 (sentence case for headings and titles) edits the five moved vendor pages and prometheus-v2.mdx, which this PR moves or deletes, plus prometheus.mdx, coralogix.mdx, how-it-works.mdx, resolve/traces/overview.mdx and resolve/overview.mdx. The new pages here already use sentence case. Whichever lands second keeps this PR's moves and deletions and re-applies docs: sentence case for headings and titles #774's casing to the headings this PR did not write.
• docs(private-locations): agent Prometheus metrics and single server port #755 (private location agent metrics) merges cleanly with this branch. Its agent-metrics.mdx links /integrations/observability/prometheus-v2/, which the new redirect covers; it should point at /integrations/observability/prometheus.