[reckless-sherixx]

Summary

Adds a page for ClickHouse logs: how to turn on collection, choose ClickHouse as the logs provider, which LogQL works, how it differs from Loki, the limits, troubleshooting, and how to turn it off. The Logging overview now points to it.

Stays a draft until chart 0.1.31 is released. Every command pins that version, and today no chart with that exact version exists.

Decision: the page documents the awkward first upgrade on clusters without ClickHouse (--no-hooks plus a password), because that's the only route that worked. If a chart fix lands first, that section shrinks to one command. Please look hardest at "Existing install without ClickHouse" and "JSON logs".

Type of change

• New documentation page
• Update to existing documentation

Checklist

• I ran npm run build locally and the site builds without errors
• I ran npm run lint locally and lint passes
• Internal links work; new external links open the correct page
• My commits are signed off (DCO — git commit -s)
• I read CONTRIBUTING.md

Additional notes

What has to happen before this leaves draft

• Release chart 0.1.31 and rerun the commands. Helm upgrades were tested with 0.1.31-rc-1 on a one-node kind cluster with no NudgeBee backend. The exact helm repo add and --version 0.1.31 commands have not been run.
• Confirm the self-hosted server version. The page says 1.15.0, which isn't released yet. Change it if the release doesn't contain the logs and Nubi work this page describes.
• See the UI flow on a real screen. The toggle, Log Provider menu, Agent Health text and error messages come from reading the code.
• Check metric queries on Query Log. The table lists them as supported, but I left them out of the examples because their behavior on that screen isn't settled.
• Test whether changing ttl affects an existing table. The page gives the default retention and doesn't say how to change it, because a change may only apply to new tables.

Not in this PR

• helm_values.md and agent-health.md: open draft docs: update pages for the latest release (agent 0.1.30) #892 already changes them, so logs.collection and the health text belong there.
• External ClickHouse: not documented until there's a supported path.
• Other pages that this work found to be wrong or risky, left for their own change:
  • upgrade.md says the collector and ClickHouse "come back" when installation.sh -a is re-run. On a cluster with no ClickHouse Secret, that likely fails with PASSWORDS ERROR.
  • loki.md, clickhouse-tracing.md and upgrade.md use --reuse-values, which keeps the old collector settings across chart versions.
  • storage-and-pvcs.md tells readers to change ttl, which may not affect an existing table.

[gemini-code-assist]

Code Review

This pull request introduces comprehensive documentation for the new ClickHouse Logs feature, which allows the NudgeBee Agent to collect container logs into its bundled ClickHouse instance. A new detailed guide (clickhouse.md) has been added to cover setup, operation, LogQL support, and troubleshooting, and the main logging index page (index.md) has been updated to reference this new feature. There are no review comments, and I have no feedback to provide.