jcoffey-dev
4f0da1ae5e
Extends the heartbeat mechanism with a second gRPC service on the same mTLS channel (AgentControl.CheckIn, agent-initiated on the existing heartbeat ticker -- still push-only, no inbound port on any agent) so an agent reports its running config and can pick up an operator-set override. A new web UI section (/agents) lists every agent that's checked in, shows its reported config, and lets an operator edit a narrow, deliberately-scoped subset remotely: batch/heartbeat tuning, and (journald sources only) the unit filter. TLS material and the ingest endpoint are never reportable or remotely editable, by proto shape rather than a validation rule -- a bad or malicious edit there could permanently strand an agent or redirect where its logs go, unlike every other editable field, which only degrades behavior. An override lives only in the agent's memory (agent.toml is never rewritten) and re-syncs on the agent's own schedule; changing the journald filter aborts and respawns the source task since there's no other way to change what's being tailed. Building the hot-reload path surfaced a real, independent, pre-existing bug: shutdown was using poll_timeout(), which only drains once flush_interval has elapsed, silently dropping anything buffered more recently on every graceful shutdown that landed between flushes -- fixed with a new unconditional Batcher::flush_all(), now used at both shutdown and hot-reload. Verified live end-to-end against a real stack: an edited heartbeat interval changed a running agent's actual send cadence within one check-in cycle (confirmed by the real timestamps landing in ClickHouse), and an edited journald filter triggered a real source restart, both reflected back in the next reported-config snapshot. See /docs/agent-management-design.md.
web
SvelteKit frontend. Phase 0: one page, one query box, one table. No auth,
no styling polish, no routing beyond /.
What it does
Textarea for a raw SQL string → POST {VITE_API_BASE_URL}/query on /api
→ renders {columns, rows} as an HTML table, or shows {error} from a
rejected/failed query. That's the whole app — see src/routes/+page.svelte.
Why a static build, not a Node server
Scaffolded with @sveltejs/adapter-static: this page has no server-side
data loading (all data comes from a client-side fetch triggered by the
submit button), so there's nothing here that needs a running SvelteKit
server. A prerendered static site is simpler to build, deploy, and reason
about than running Node in production for a page that's this thin.
Because it's static, VITE_API_BASE_URL is baked in at build time, not
read at container start. Set it before npm run build (or pass
--build-arg VITE_API_BASE_URL=... to docker build) — changing it later
means rebuilding, not just restarting the container.
Building & running
npm install
cp .env.example .env # adjust VITE_API_BASE_URL if /api isn't on localhost:8080
npm run dev # local dev server with hot reload
npm run check # svelte-check, type errors
npm run build # static output to build/
npm run preview # serve the static build locally to sanity-check it
docker build -f Dockerfile -t sentry-web . # context is web/, not the repo root
docker run -p 3000:3000 sentry-web
Tenant picker (Phase 4)
src/routes/select-tenant is the one route that isn't reachable by
clicking around the app -- enterprise-auth's internal/loginhandler
redirects a browser here after an SSO login resolves to more than one
tenant_memberships row (see that package's doc comment), carrying a
short-lived sentry_pending_login cookie instead of a real session. The
page calls GET /auth/memberships to list the choices, and
POST /auth/select-tenant on a click, both via
fetch(..., {credentials: 'include'}) ($lib/api.ts's
listMemberships/selectTenant) so that cookie -- and, on success, the
real session cookie the POST response sets -- actually cross the origin
boundary between this app and enterprise-auth. enterprise-auth's
default POST_LOGIN_REDIRECT_URL (this app's own base URL) is also
where CORS_ALLOWED_ORIGIN defaults to, and it has to be a literal
origin, not * -- see api/httpserver.WithCredentialedCORS's doc
comment for why a credentialed fetch and a wildcard CORS origin can
never be combined; getAuthFeatures above deliberately doesn't send
credentials for exactly this reason, and is why it could stay on the
plain WithCORS every other endpoint in this repo uses.
Like every other route (export const prerender = true in this route's
own +page.ts), no server-side data loading -- the membership list and
the tenant choice both come from client-side fetch calls the same way
the root query page's does.
Verified in a real browser in this environment: a throwaway Node server
standing in for enterprise-auth (implementing the exact
GET /auth/memberships/POST /auth/select-tenant wire contract,
including the plain-text http.Error bodies the real handler sends, not
JSON) on a different origin/port than this app's dev server, driven
through the full flow -- cross-origin pending-login cookie set, the
credentialed preflight + GET/POST round trip, a real click choosing
a tenant, and the post-selection redirect landing back on / -- plus
the missing/expired-cookie error path separately. No Docker or live
Postgres/IdP needed for this, since the whole point was exercising this
app's own fetch/CORS/cookie wiring against a contract-accurate fake, not
enterprise-auth's internals (those are enterprise/internal/ loginhandler's own tests' job, already covered there).
Why nginx, not distroless
The repo convention prefers distroless/scratch base images. Serving a
static SPA still needs some HTTP server, though, and nginx:alpine is
the boring, standard choice for that job — writing a custom static-file
binary just to stay distroless would be more engineering than a Phase 0
placeholder page justifies. nginx.conf here is minimal: serve build/,
fall back to index.html for client-side routing (only one route exists
today, but this is what you want the moment a second one is added).