mirror of
https://github.com/logos-messaging/logos-delivery.git
synced 2026-07-31 08:53:12 +00:00
6 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Igor Sirotin
|
8125cd0262
|
refactor(metrics): give every metric a logos_delivery_ prefix (#4074)
* refactor(metrics): prefix node metrics with logos_delivery_
Every metric the node exports now starts with logos_delivery_. There was no
prefix mechanism before: nim-metrics derives the exported name from the Nim
identifier, no declaration passed an explicit `name = "..."`, and the waku_
convention was maintained by hand -- 69 of the 80 node metrics followed it and
11 did not (query_count, query_time_secs, event_loop_load,
event_loop_accumulated_lag_secs, postgres_payload_size_bytes, reconciliation_*,
total_* and the camelCase rendezvousPeerFoundTotal).
Identifiers are renamed rather than given a `name = "..."` argument, keeping the
invariant that the Nim identifier is the exported name and letting the compiler
check every call site.
rendezvousPeerFoundTotal becomes logos_delivery_rendezvous_peer_found: it was the
only camelCase metric, and the trailing Total was redundant since nim-metrics
already appends _total to counters at exposition time.
library/ is untouched on purpose -- `proc waku_version()` in
kernel_api/debug_node_api.nim is the exported libwaku C ABI symbol, not the gauge
of the same name in node_telemetry.nim.
BREAKING CHANGE: metric names change. Dashboards, alert rules and recording rules
that reference waku_* must be updated; see docs/operators/how-to/monitor.md.
* refactor(metrics): prefix auxiliary app metrics with logos_delivery_
Applies the same prefix to the tools shipped from this repo: liteprotocoltester
(lpt_*), networkmonitor (networkmonitor_*), chat2bridge (chat2_*) and the
lightpush_mix example (lp_mix_*).
These tools are not the delivery node and already had their own consistent
prefixes, so this commit is separable from the node rename if the intent was to
namespace only the node itself.
* chore(metrics): query old and new metric names in Grafana dashboards
212 expressions across 9 dashboards now match both the waku_* and the
logos_delivery_* spelling, so panels keep working across the upgrade and over
historical data:
sum by (type)((increase(waku_node_errors_total{...}[$__rate_interval])
or increase(logos_delivery_node_errors_total{...}[$__rate_interval])))
The `or` is placed around the leaf, inside every aggregation. That depth is
load-bearing: `or` keeps its right operand only for label sets absent from the
left, so `sum by (type)(old) or sum by (type)(new)` aggregates each half of the
fleet separately and then discards the right one entirely -- silently dropping
every already-upgraded node. 36 panels here collapse `instance`.
Measured against a local Prometheus scraping two targets, one exporting old names
at 10/s and one exporting new names at 20/s (truth 30/s): union outside the
aggregation gives 10, union around the leaf gives 30.
Where the leaf sits in a range vector the whole call is duplicated, since
`(a or b)[5m]` is not valid PromQL.
Once every scraped node runs a release with the new names and the old samples
have aged out of retention, the `or` half can be deleted.
* test(e2e): expect logos_delivery_-prefixed metric names
The e2e suite asserts against a live /metrics endpoint, which serves only the new
names, so these are replaced rather than unioned. libp2p_* entries are unchanged.
* docs(operators): document the logos_delivery_ metric prefix
Records that every metric the node exports is prefixed, that dependency metrics
(libp2p_*, nim_gc_*, process_*) keep their own names, and shows where the `or`
has to sit if operators maintain their own dashboards or alert rules.
* refactor(metrics): name the store fleet metrics after store, not relay
logos_delivery_relay_fleet_store_msg_size_bytes and _msg_count are declared in
waku_store/protocol_metrics.nim and recorded by the store client, but carried a
relay prefix. Renamed to logos_delivery_store_fleet_msg_size_bytes and
logos_delivery_store_fleet_msg_count.
The dashboard keeps matching the old exported name, which was
waku_relay_fleet_store_*.
Note that both metrics are wrong independently of their name, see the PR
description.
|
||
|
Fabiana Cecin
|
ce918b0819
|
chore: replace Option with Opt (#4035)
* change all usage of std.options.Option[T] to results.Opt[T] * fix broken apps and examples (to validate refactor) * removed all std/options code added for libp2p v2 migration * add broker Opt codec to persistency/backend_comm.nim * add a readValue overload for Opt[T] in tools/confutils/cli_args.nim * keep Option where required (Presto, confutils' config_file.nim) * Change generateRlnProof error handling * Fix imports |
||
|
Ivan FB
|
0a19473a3b
|
fix: don't apply reconnect backoff to discovered relay peers (#4029) | ||
|
NagyZoltanPeter
|
a7f893555d
|
Integrate api-shape phase2 (#3989) + api interfaces (#3975) (#3999)
* Reshape per-layer API into api/ folders and thin the FFI over them
Each layer now separates its constructible core from its public surface:
- core module (waku.nim / messaging_client.nim /
reliable_channel_manager.nim): the type plus new/start/stop and the
private construction helpers.
- api/ folder: one module per differentiated set of operations
(waku: topics/relay/filter/lightpush/store/peer_manager/discovery/
debug/health) plus an events surface.
The waku api is reshaped to be the complete operation surface the C
bindings need, so the library no longer reaches into node internals:
relayPublish returns the message hash, relaySubscribe takes an optional
handler, filter/lightpush auto-select the service peer, connectedPeersInfo
returns structured data, pingPeer honours the timeout, plus
relayNumPeersInMesh / relayNumConnectedPeers / isOnline. library/ is now a
thin C-ABI shim: each {.ffi.} proc only marshals cstring/JSON/callbacks and
delegates to ctx.myLib[].waku.<op> (or messagingClient.<op>).
app_callbacks re-exports the modules defining its handler types, which the
included FFI files previously relied on by leakage.
Events move next to the surface that owns them, with each dependency kept
pointing the right way:
- waku/events/ relocated under waku/api/events/.
- channel events live in channels/api/events.nim.
- the four messaging-level message events move to messaging/api/events;
MessageSeenEvent stays in waku because it is emitted by waku core, so
moving it would make waku depend on the messaging layer.
- delivery_events renamed to filter_subscribe_events to match the
OnFilterSubscribe/Unsubscribe events it actually declares.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Add reliable-channel FFI ops + events (nim-ffi v0.1.3)
Expose the reliable-channel layer through the v0.1.3 FFI:
- channel_create / channel_send / channel_close call the
ReliableChannelManager api (createReliableChannel / send / closeChannel),
marshalling channel id + base64 payload + ephemeral by hand
- channel message received / sent / errored are surfaced by listening to the
channel-layer broker events in start_node and forwarding them through
callEventCallback (received payload base64-encoded), dropped in stop_node
Stays on nim-ffi v0.1.3 (no typed/CBOR rewrite).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Expose reliable-channel ops in the stable C header (#3851)
The library already ships as a single .so with a tiered header surface
(liblogosdelivery.h = stable Messaging/Reliable-Channels, liblogosdelivery_kernel.h
= advanced Kernel). Per that tiering, the reliable-channel ops belong on the
stable surface, so declare channel_create / channel_send / channel_close in
liblogosdelivery.h and document the channel lifecycle events delivered through
the event callback.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Graft PR#3975 interface layer onto decomposed foundation (events deduped)
Add IKernel/IMessagingClient/IReliableChannelManager/ILogosDelivery interface
classes under logos_delivery/api/. The EventBroker types PR#3975 hoisted into
these files already exist in PR#3989's decomposed */api/events/ modules, so the
interface files re-export those modules instead of redefining the types
(avoids 8 duplicate EventBroker definitions). api/types.nim kept at the
foundation version (ChannelId stays in channels/types.nim, which the decomposed
modules import).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Wire impl classes to interfaces (inherit; relocate SendHandler)
- Waku : IKernel, MessagingClient : IMessagingClient,
ReliableChannelManager : IReliableChannelManager.
- The operation procs already live in PR#3989's decomposed */api/ modules and
stay as plain procs (nothing dispatches through the interface types, so no
method-ization is needed).
- SendHandler now lives in reliable_channel_manager_api.nim (its PR#3975 home);
removed the duplicate from reliable_channel.nim, which re-exports the
interface module so channels/api/{channel_lifecycle,send} still see it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Wire LogosDelivery to ILogosDelivery orchestrator interface
LogosDelivery : ILogosDelivery; start/stop/isOnline become method overrides.
Peripheral PR#3975 edits (lightpush/store clients, self_req_handlers,
statistics) are import-reorg artifacts of deleting waku/utils/requests.nim,
which the decomposed structure keeps -- so they are intentionally not ported.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Dedup EventConnectionStatusChange (re-export from health_events)
9th duplicate EventBroker type: defined in both logos_delivery_api.nim and the
decomposed waku/api/events/health_events.nim. The interface file now re-exports
it. liblogosdelivery builds clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Move events back into interface-class source files (restore #3975 placement)
Reverses the earlier dedup-by-re-export: event TYPE definitions now live in the
interface classes, and the emptied decomposed event files are removed.
- MessageSeenEvent -> logos_delivery/api/kernel_api.nim
- Message{Sent,Error,Propagated,Received}Event -> api/messaging_client_api.nim
- ChannelMessage{Received,Sent,Error}Event -> api/reliable_channel_manager_api.nim
- EventConnectionStatusChange -> api/logos_delivery_api.nim
Deleted (became empty after the move):
- logos_delivery/waku/api/events/message_events.nim
- logos_delivery/messaging/api/events.nim
- logos_delivery/channels/api/events.nim
health_events.nim keeps its two remaining events (content/shard topic health).
Rewiring: each layer re-exports its interface module (waku->kernel_api,
messaging_client->messaging_client_api, reliable_channel->reliable_channel_manager_api,
which also re-exports messaging_client_api). Deep emitters/listeners
(subscription_manager, waku_node, waku_node/relay, node_health_monitor,
recv_service, send_service) import the owning interface module directly.
kernel_api stays below node level (types/topics/message/store-common) so the
node->kernel_api imports are acyclic. liblogosdelivery builds.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* nph formatting
---------
Co-authored-by: Ivan FB <ivansete@status.im>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
|
||
|
Fabiana Cecin
|
6837ae0c1f
|
feat: bump nim-libp2p to v2.0.0 (#3929)
* bump nim-libp2p pin to v2.0.0 tag * bump json_rpc to v0.6.1, lsquic to v0.5.1, boringssl to v0.0.8 (latest tags) * add libp2p_mix dep; repoint libp2p/protocols/mix -> libp2p_mix * pin nimble.lock: websock / protobuf_serialization / npeg / jwt * Makefile: add -d:libp2p_quic_support * regenerate nix/deps.nix (adds libp2p_mix, refreshes pins) * migrate rng ref HmacDrbgContext -> libp2p Rng across prod/channels/tests (interface-only; same DRBG) * waku_switch: TransportConfig factory; unified 2.0.0 connection limits (withMaxInOut, withMaxConnections); local MaxConnections * waku_relay/rendezvous/discv5/kademlia: v2.0.0 API (rng, config, ServiceDiscovery rename) * call Service.setup() on post-build switch services (2.0.0 split setup/start) * drop libp2p/utils/semaphore -> chronos AsyncSemaphore * add logos_delivery/waku/compat/option_valueor shim (Option[T] valueOr/withValue, dropped upstream) * add std/options where a transitive re-export was removed * add newStandardSwitch shim (libp2p removed it in 2.0.0); mounts yamux+mplex to match prod muxer * PeerId.random(rng); common.rng()/crypto.newRng(); hoist shared rng (instantiation cleanup) * update expectations for 2.0.0 defaults: DEFAULT_PROTOCOLS += /ipfs/id/push/1.0.0; agent "nim-libp2p" * drop relay reboot/reconnect test (asserted a Switch restart capability that is simply not supported) * fix up a few tests that were flaking on MacOS (libp2p upgrade may have exposed these) |
||
|
Ivan FB
|
3b03ca29b1
|
refactor: introduce proper logos_delivery layers folder structure (#3935)
Co-authored-by: NagyZoltanPeter <113987313+NagyZoltanPeter@users.noreply.github.com> |