Ilia Denisov
bb71e7b1c7
CI / changes (pull_request) Successful in 2s
CI / unit (pull_request) Successful in 12s
CI / integration (pull_request) Successful in 26s
CI / ui (pull_request) Successful in 1m11s
CI / conformance (pull_request) Successful in 10s
CI / gate (pull_request) Successful in 0s
CI / deploy (pull_request) Successful in 1m56s
The bot runs on its own host and exports no telemetry (otelcol is unreachable
from there), so it was a monitoring blind spot. Observe its Bot API health
centrally by wrapping the HTTP client — one place, no per-call-site
instrumentation — and relay it up the existing bot-link as a periodic Health
message the gateway turns into its own metrics.
- proto: add Health (delta connect / api / 429 counters + a last-ok stamp) to
the FromBot oneof (additive, backward-compatible).
- bot: platform/telegram/internal/health wraps the Bot API HTTP client — it
stamps liveness on any 2xx (so the getUpdates long-poll keeps it fresh even
when idle), classifies transport/5xx failures (getUpdates vs other) and 429s,
and honours a 429's Retry-After (bounded) so the bot backs off; a 4xx other
than 429 is a normal per-request outcome and is not counted.
- bot-link client: flush the reporter as a Health message every 30s over a
single-sender loop (a gRPC stream forbids concurrent Send).
- gateway: fold each report into bot_tg_errors_total{kind} and the
bot_tg_last_ok_unix liveness gauge.
- grafana: alerts (bot disconnected, bot not reaching the Bot API, sustained
429s) routed to the operator email — which does not go through the bot — plus
a Telegram-bot dashboard.
- docs (ARCHITECTURE, compose comments); unit tests (observer classification,
Retry-After, snapshot/commit, hub last-ok monotonicity).
pkg
Shared wire contracts for the Scrabble platform (module scrabble/pkg),
imported by both backend and gateway. It carries no logic — only the
generated message types and the schemas they come from.
Layout
proto/push/v1/ # backend -> gateway live-event gRPC channel (Push.Subscribe)
# committed generated Go (*.pb.go, *_grpc.pb.go)
fbs/scrabble.fbs # FlatBuffers edge payloads (one `scrabblefb` namespace)
fbs/scrabblefb/ # committed generated Go for the schema
proto/push/v1is the single gRPC server-stream the backend exposes and the gateway subscribes to (Event{user_id, kind, payload, event_id}); thepayloadis an opaque FlatBuffers body the gateway forwards verbatim.proto/telegram/v1is the Telegram connector's RPC contract (includingValidateLoginWidgetfor the web Login Widget sign-in).fbsholds the client↔gateway request/response and event payloads as FlatBuffers tables. The backend encodes the push payloads from these types; the gateway transcodes the rest to and from the backend's JSON; the UI generates TypeScript from the same.fbs.
Generated code
Committed (CI only builds it); regenerate dev-time after editing the schemas:
make -C pkg tools # go install protoc-gen-go + protoc-gen-go-grpc
make -C pkg gen # buf generate (proto) + flatc (fbs)
flatc is pinned to 23.5.26 to match the github.com/google/flatbuffers
Go runtime in go.mod; generating with another version is refused.
Workspace wiring
scrabble/pkg is a bare-path module (no dot), so — like scrabble-solver — it
cannot be fetched as a versioned dependency. go.work carries use ./pkg and
replace scrabble/pkg v0.0.0 => ./pkg; consumers require scrabble/pkg v0.0.0.