* client/network: Re-enable light_client_handler.rs unit tests
* client/network: Add scaffolding for light client using req-resp
* client/network: Make it compile
* client/network: Rename OutEvent SendRequest
* client/network: Restructure light client request client and handler
* client/network: Rename light client request client to sender
* client/network: Remove light client prepare_request
* client/network/src/light: Rework configuration
* client/network: Formatting
* client/network/light: Remove RequestId
* client/network/light: Make request functions methods
* client/network/light: Refactor request wrapping
* client/network/light: Fix warnings
* client/network/light: Serialize request in method
* client/network/light: Make returning response a method
* client/network/light: Depend on request response to timeout requests
* client/network: Fix test compilation
* client/network/light: Re-enable connection test
* client/network/light: Re-enable timeout test
* client/network/light: Re-enable incorrect_response test
* client/network/light: Re-enable wrong_response_type test
* client/network/light: Re-enable retry_count_failures test
* client/network/light: Re-enable issue_request tests
* client/network/light: Re-enable send_receive tests
* client/network/light: Deduplicate test logic
* client/network/light: Remove unused imports
* client/network/light: Handle request failure
* client/network/light: Move generate_protocol_config
* client/network: Fix test compilation
* client/network: Rename light client request client to sender
* client/network: Handle too-many-requests error
* client/network: Update outdated comments
* client/network/light: Choose any peer if none has best block defined
* .maintain: Replace sentry-node with local-docker-test-network
Sentry nodes are deprecated. Thus there is no need for
`.maintain/sentry-node` to spin up a sentry node test environment.
Instead this commit rewrites the setup to contain two full-connected
validators and one light client.
With the steps below one can now spin up a local test network with
two validators, one light-client, Prometheus and Grafana.
- cargo build --release
- sudo docker-compose -f .maintain/local-docker-test-network/docker-compose.yml up
* client/network/light: Handle oneshot cancellation
* client/network/light: Do not reduce retry count on missing peer
* client/network/request-response: Assert in debug request id to be unique
* client/network/light: Choose same limit as block request protocol
* client/network: Report reputation changes via response
Allow request response protocol handlers to issue reputation changes, by
sending them back along with the response payload.
* client/network: Remove resolved TODOs
* client/network: Report reputation changes via response
When handling a request by a remote peer in a request response handler,
one might want to in- or de-crease the reputation of the peer. E.g. one
might want to decrease the reputation slightly for each request, given
that it forces the local node to do work, or one might want to issue a
larger reputation change due to a malformed request by the remote peer.
Instead of having to pass a peerset handle to each request response
handler, this commit suggests to allow handlers to isssue reputation
changes via the provided `pending_response` `oneshot` channel.
A reputation change issued by a request response handler via the
`pending_response` channel is received by the
`RequestResponsesBehaviour` which passes the reputation change up as an
event to eventually be send to a peerset via a peerset handle.
* client/network/req-resp: Use Vec::new instead of None::<Vec<_>>
* client/network: Rename Response to OutgoingResponse
Given that a request-response request is not called `Request` but
`InomingRequest`, rename a request-response response to
`OutgoingResponse`.
* client/finality-grandpa-warp: Send empty rep change via response
* client/network/req-resp: Add unit test for request id collision
* client/network/req-resp: Prevent request id collision
`RequestId` is a monotonically increasing integer, starting at
`1`. A `RequestId` is unique for a single `RequestResponse`
behaviour, but not across multiple `RequestResponse` behaviours. Thus
when handling `RequestId` in the context of multiple
`RequestResponse` behaviours, one needs to couple the protocol name
with the `RequestId` to get a unique request identifier.
This commit ensures that pending requests (`pending_requests`) and
pending responses (`pending_response_arrival_time`) are tracked both by
their protocol name and `RequestId`.
* client/network/req-resp: Remove unused import
* client/network/req-resp: Introduce ProtocolRequestId struct
* client/network/req-resp: Update test doc comment
Treat `RequestId` as an opaque type.
* client/network/req-resp: Improve expect proof
Sentry nodes are deprecated. Thus there is no need for
`.maintain/sentry-node` to spin up a sentry node test environment.
Instead this commit rewrites the setup to contain two full-connected
validators and one light client.
With the steps below one can now spin up a local test network with
two validators, one light-client, Prometheus and Grafana.
- cargo build --release
- sudo docker-compose -f .maintain/local-docker-test-network/docker-compose.yml up
* *: Update to libp2p v0.33.0
* client/network: Consistently track request arrival time
With https://github.com/libp2p/rust-libp2p/pull/1886/ one is guaranteed
to receive either a `ResponseSent` or a `InboundFailure` event for each
received inbound request via `RequestResponseEvent::Message`. Given this
guarantee there is no need to track arrival times in a best-effort
manner and thus there is no need to use a LRU cache for arrival times.
* client/offchain: Adjust to PeerId API changes
* *: Update to libp2p v0.32.0
* Cargo.lock: Update async-tls to 0.10.2
* client/network/request_response: Adjust to new request response events
* client/network/request_response.rs: Clean up silently failing responses
* client/network/discovery: Lazily instantiate mdns
* client/network/discovery: Exclude MdnsWrapper for target_os unknown
* client/network/discovery: Fix indentation
* client/network/request-response: Use LruCache to track pending resp time
* client/network/request_responses: Fix early connection closed error
* client/network/request-response: Replace debug_assert with debug
* client/network/request-response: Fix typo
* client/network/request-response: Don't emit event on send_response fail
* client/network/request-response: Revert waker.wake_by_ref()
* client/network/request-resp: Make duration in InboundRequest optional
* client/network/req-resp: Don't emit two events for busy builder
When a response builder is busy incoming requests are dropped.
Previously this was reported both via a `ResponseFailure::Busy` and a
`ReponseFailure::Network(InboundFailure::Omisssion)` event.
With this commit the former is removed, leaving only the latter in
place.
* *: Update authority discovery and remove WorkerConfig
With https://github.com/paritytech/substrate/pull/7545 the authority
discovery module queries and publishes addresses on an exponentially
increasing interval. Doing so should make custom configurations
obsolete, as operations are retried in a timely fashion in the first
minutes.
* */Cargo.{lock,toml}: Point to mxinden substrate auth-disc-timing
* Revert "*/Cargo.{lock,toml}: Point to mxinden substrate auth-disc-timing"
This reverts commit 0785943a1e377454f088814ef20f4432de09da7a.
* "Update Substrate"
* Revert ""Update Substrate""
This reverts commit 377b221e1853b2c383f0c416d686535b545796cb.
* Cargo.lock: Manual Substrate update
* node/test/service/src/lib: Remove unused import
* parachain/test-parachains/adder: Remove unused import
Co-authored-by: parity-processbot <>
* client/authority-discovery: Publish and query on exponential interval
When a node starts up publishing and querying might fail due to various
reasons, for example due to being not yet fully bootstrapped on the DHT.
Thus one should retry rather sooner than later. On the other hand, a
long running node is likely well connected and thus timely retries are
not needed. For this reasoning use an exponentially increasing interval
for `publish_interval`, `query_interval` and
`priority_group_set_interval` instead of a constant interval.
* client/authority-discovery/src/interval.rs: Add license header
* .maintain/gitlab: Ensure adder collator tests are run on CI
* node/service/src/lib: Do not spawn authority discovery on sentries
The notion of sentry nodes has been deprecated (see [1] for details).
Support for sentry nodes in the `client/authority-discovery` module has
been removed.
This commit adjusts the instantiation of the authority discovery worker
accordingly, only spawning the module on authority nodes.
[1] https://github.com/paritytech/substrate/issues/6845
* "Update Substrate"
Co-authored-by: parity-processbot <>
* client/authority-discovery: Remove sentry node logic
The notion of sentry nodes has been deprecated (see [1] for details).
This commit removes support for sentry nodes in the
`client/authority-discovery` module.
While removing `Role::Sentry` this commit also introduces
`Role::Discover`, allowing a node to discover addresses of authorities
without publishing ones own addresses. This will be needed in Polkadot
for collator nodes.
[1] https://github.com/paritytech/substrate/issues/6845
* client/authority-discovery/service: Improve PeerId comment
As done with `NetworkService::{add_to,remove_from}_priority_group`, make
`NetworkService::set_priority_group` async as well. This future-proofs
the API should we ever decide to use a bounded channel between
`NetworkService` and `NetworkWorker`.
The Rust libp2p-kad implementation can require iterative queries to use
disjoint paths for increased resiliency in the presence of potentially
adversarial nodes.
Allow Substrate users to enable this feature via the
`--kademlia-disjoint-query-paths` flag.
* *: Bump async-std to v1.6.5
Prevent users from using v1.6.4 which faces issues receiving incoming
TCP connections. See https://github.com/async-rs/async-std/issues/888
for details.
* client/network/src/gossip: Use channel instead of condvar
`async_std::sync::Condvar::wait_timeout` uses
`gloo_timers::callback::Timeout` when compiled for
`wasm32-unknown-unknown`. This timeout implementation does not fulfill
the requirement of being `Send`.
Instead of using a `Condvar` use a `futures::channel::mpsc` to signal
progress from the `QueuedSender` to the background `Future`.
* client/network/Cargo.toml: Remove async-std unstable feature
* client/network/src/gossip: Forward all queued messages
* client/network/gossip: Have QueuedSender methods take &mut self
* client/network/gossip: Move queue_size_limit into QueuedSender
The `queue_size_limit` field is only accessed by `QueuedSender`, thus
there is no need to share it between the background future and the
`QueuedSender`.
* client/network/gossip: Rename background task to future
To be a bit picky the background task is not a task in the sense of an
asynchonous task, but rather a background future in the sense of
`futures::future::Future`.
With the `OnRead` flow control option yamux "send[s] window updates only
when data is read on the receiving end" and not as soon as "a Stream's
receive window drops to 0".
Yamux flow control has proven itself. This commit removes the feature
flag. Yamux flow control is now always enabled.
This reverts commit a7b6c91f80.
The authority discovery module was initially enabled by default on
validator and sentry nodes with commit 59f5eb4. This change was later on
reverted in a7b6c91. With this commit the authority discovery module is
again enabled by default.
* .maintain/monitoring: Add alert when continuous task ends
Through the `polkadot_tasks_ended_total` Prometheus metric one can tell
when a task ended. Use this metric to alert when specific
known-to-be-continuous tasks end on a node.
* .maintain/monitoring: Don't hard-code task names
* .maintain/monitoring: Normalize alerting rules
- Start alert names with their component and end with the describing
adjective.
- Describe alert duration in `message` with `for more than` across all
alerts.
* .maintain/monitoring: Fix alert tests
Stop sending legacy handshake payloads, i.e. length-prefixed protobuf
payloads inside a length-prefixed noise frame. Receiving of such legacy
handshakes is still supported, thus backwards compatibility is given.
Instead of building a libp2p `Transport` via a chain of `and_then`,
leverage the `libp2p::core::upgrade::Builder` pattern for authentication
and multiplexing.
Extend `sub_libp2p_kbuckets_num_nodes` Prometheus metric to expose the
number of nodes per bucket per Kademlia instance instead of only per
Kademlia instance.
* client/authority-discovery: Throttle DHT requests
Instead of passing one DHT query for each authority down to the network
every query interval, only pass MAX_IN_FLIGHT_LOOKUPS at a given point
in time, triggering new ones when previous ones return.
* client/authority-discovery/worker/test: Fix wrong constant
* frame/authority-discovery: Have authorities() return both current and next
Authority address lookups on the DHT happen periodically (every 10
mintues) and are rather slow (~10 seconds).
In order to smooth the transition period between two sessions, have the
runtime module return both the current as well as the next authority
set. Thereby the client authority module will:
1. Publish its addresses one session in advance.
2. Prefetch the addresses of authorities of the next session in advance.
* frame/authority-discovery: Deduplicate authority ids
* frame/authority-discovery: Don't dedup on_genesis authorities
* frame/authority-discovery: Remove mut and sort on comparison in tests
* frame/authority-discovery: Use BTreeSet for deduplication
* client/cli/src/config: Warn on low file descriptor limit
Substrate sets the soft file descriptor limit to the hard limit at
startup. In the case of the latter being low already (< 10_000) a
Substrate node under high demand might run into issues e.g. when opening
up new TCP connections or persisting data to the database.
With this commit a warn message is printed to stderr.
* client/cli/Cargo.toml: Update to fdlimit 0.2.0
* client/*: Treat protocol name as str and not [u8]
Notification protocol names are in practice always valid utf8 strings.
Instead of treating them as such in the type system, thus far they were
casted to a [u8] at creation time.
With this commit protocol names are instead treated as valid utf8
strings throughout the codebase and passed as `Cow<'static, str>`
instead of `Cow<'static, [u8]>`. Among other things this eliminates the
need for string casting when logging.
* client/network: Don't allocate when protocol name is borrowed
* client/authority-discovery/worker: Extract address getter
* client/authority-discovery: Test for no duplicate p2p components
* client/authority-discovery: Append PeerId to Multiaddr at most once
When collecting the addresses to be published for the local node,
`addresses_to_publish` adds the local nodes `PeerId` to each
`Multiaddr`. Before doing so, ensure the `Multiaddr` does not already
contain one.
* client/authority-discovery: Remove explicit return
* *: Update to next libp2p version (likely v0.24.0)
* Revert "*: Update to next libp2p version (likely v0.24.0)"
This reverts commit ffe1545aba6c2557a2843579de331f3fc1c60743.
* */Cargo.toml: Update to libp2p v0.24.0
* client/network/src/service: Handle ConnectionClosed returning Option
* Cargo.*: Test kad usize conversion
* Revert "Cargo.*: Test kad usize conversion"
This reverts commit ad317879782f982cb4a4c76029a72b5b97e82bec.
* Cargo.lock: Update to libp2p-kad v0.22.1
* client/cli/Cargo.toml: Update to libp2p 0.24.0
* service/src/lib: Update authority discovery construction
https://github.com/paritytech/substrate/pull/6760 introduces the concept
of an authority discovery `Service` allowing one to communicate with an
authority discovery `Worker`, e.g. to learn the `Multiaddr`s for a given
`AuthorityId`.
Along with the new `Service` structure it also alters the authority
discovery constructor to return both a worker and a service. This
commit adjusts the callside of the constructor, ignoring the `Service`
for now.
* "Update Substrate"
* Revert ""Update Substrate""
This reverts commit 04fb79c465f91b55422e22d4ea266f08f4072854.
* Update Substrate
Co-authored-by: parity-processbot <>
* client/authority-discovery: Rename AuthorityDiscovery to XXXWorker
* client/authority-discovery: Introduce AuthorityDiscoveryService
Add a basic `AuthorityDiscoveryService` implementation which enables
callers to get the addresses for a given `AuthorityId` from the local
cache.
* client/authority-discovery: Split into worker and service mod
Move `Service` and `Worker` to their own Rust modules resulting in the
following file structure.
├── build.rs
├── Cargo.toml
└── src
├── error.rs
├── lib.rs
├── service.rs
├── tests.rs
├── worker
│ ├── addr_cache.rs
│ ├── schema
│ │ └── dht.proto
│ └── tests.rs
└── worker.rs
* client/authority-discovery: Cache PeerId -> AuthorityId mapping
* client/authority-discovery: Update priority group on interval
Instead of updating the authority discovery peerset priority group each
time a new DHT value is found, update it regularly on an interval.
This removes the need for deterministic random selection. Instead of
trying to return a random stable set of `Multiaddr`s, the `AddrCache`
now returns a random set on each call.
* client/authority-discovery: Implement Service::get_authority_id
* client/authority-discovery: Use HashMap instead of BTreeMap
* client/authority-discovery: Rework priority group interval
* client/authority-discovery: Fix comment
* bin/node/cli: Update authority discovery constructor
* client/authority-discovery: Fuse from_service receiver
* client/authority-discovery: Remove Rng import
* client/authority-discovery: Ignore Multiaddr without PeerId
* client/authority-discovery/service: Add note on returned None
* client/authority-discovery/addr_cache: Replace double clone with deref
* client/network/src/discovery: Adjust to Kademlia API changes
* client/network: Add peers to DHT only if protocols match
With https://github.com/libp2p/rust-libp2p/pull/1628 rust-libp2p allows
manually controlling which peers are inserted into the routing table.
Instead of adding each peer to the routing table automatically, insert
them only if they support the local nodes protocol id (e.g. `dot`)
retrieved via the `identify` behaviour.
For now this works around
https://github.com/libp2p/rust-libp2p/issues/1611. In the future one
might add more requirements. For example one might try to exclude
light-clients.
* Cargo.toml: Remove crates.io patch for libp2p
* client/network/src/behaviour: Adjust to PeerInfo name change
* client/network/src/discovery: Rework Kademlia event matching
* client/network/discovery: Add trace on adding peer to DHT
* client/network/discovery: Retrieve protocol name from kad behaviour
* client/network/discovery: Fix formatting
* client/network: Change DiscoveryBehaviour::add_self_reported signature
* client/network: Document manual insertion strategy
* client/network/discovery: Remove TODO for ignoring DHT address
Co-authored-by: Pierre Krieger <pierre.krieger1708@gmail.com>
The `HighCPUUsage` alert is based on the `cpu_usage_percentage` metric.
Instead of exposing the overall CPU usage in percent, the metric exposes
the per core usage summed over all cores.
This commit removes the alert for two reasons:
1. Substrate itself does not expose the core count and thus one can not
alert based on the `cpu_usage_percentage` metric.
2. Alerting based on CPU usage is generic and not specific to Substrate
or Blockchains. Thus any CPU usage alert suffice.
* *: Enable authority discovery by default
Instead of having to explicitly enable the authority discovery module on
validator and sentry nodes, this commit enables the module by default.
Today there is no way for non validator or sentry nodes to run the
module. That might change in the future.
* service/src/lib: Fix typo in new_full! for test
Information retrieved via `DebugInfoBehaviour` is not only used for
debugging purposes, e.g. disconnecting from nodes not responding to
pings, using external addresses retrieved via indentify, ...
In order for the name to reflect the usage of the module this commit
renames it.
The transaction queue size alert has been firing with a constant 10
transactions in the queue. While maybe problematic those 10 transactions
don't need to be the same across scrape intervals.
Instead of alerting with a size above 10, alert based on two things:
1. Monotonically increasing queue size
2. Upper limit queue size reached
* client/network/service: Add primary dimension to connection metrics
Two nodes can be interconnected via one or more connections. The first
of those connections is called the primary connection.
This commit adds another dimension to the
`sub_libp2p_connections_{closed,opened}_total` metrics to differentiate
primary and non-primary connections being opened / closed.
By intuition more than one connection between two nodes is rare.
Tracking the fact whether a connection is primary or not will help prove
or disprove this intuition.
* .maintain/monitoring: Ensure to sum over all connections_closed variants
* client/network/service: Rename is_primary to is_first
* client/network/service: Split by metric name with two additional metrics
* Revert ".maintain/monitoring: Ensure to sum over all connections_closed variants"
This reverts commit 2d2f93e414440b9fc9e8f7fae6fe48bd95af6b8f.
* client/network/service: Remove labels from distinct metrics
In order to tell whether an address is the local nodes address the
authority discovery module previously compared the Multihash within the
`p2p` Multiaddr protocol.
rust-libp2p recently switched to a new PeerId representation (see [1]).
Multihashes of the same PeerId in the new and the old format don't
equal.
Instead of comparing the Multihashes, this patch ensures the module
compares the PeerIds
[1] https://github.com/libp2p/rust-libp2p/issues/555
* .gitlab-ci.yml: Run promtool on Prometheus alerting rules
Add a CI stage to test the Prometheus alerting rules within
`.maintain/monitoring`.
* .gitlab-ci.yml: Switch Prometheus stage to paritytech/tools image
* .gitlab-ci.yml: Follow http redirects in Prometheus stage
* .gitlab-ci.yml: Fix Prometheus stage promtool folder name
* client/authority-discovery: Don't add own address to priority group
In the scenario of a validator publishing the address of its sentry node
to the DHT, said sentry node should not add its own Multiaddr to the
peerset "authority" priority group.
Related to 273f31b.
* client/authority-discovery: Remove unused import PeerId
* client/authority-discovery/tests: Add tcp protocol to multiaddresses
Create a place to collaborate on Prometheus alerting rules for
Substrate starting with a basic set of rules covering:
- Resource usage
- Block production
- Block finalization
- Transaction queue
- Networking
- ... Others
Substrate was previously running both a chain specific Kademlia DHT as
well as the `/ipfs/kad/1.0.0` Kademlia DHT for backwards compatibility.
(See 4db45a8 for details.)
With the end of the transition period this commit removes the former
only leaving the latter and thus preventing the discovery of
incompatible nodes via Kademlia.
An `UntilImported` stream wraps a `Stream` of incoming messages and
waits for blocks those messages are based on before passing the messages
on.
The above `Stream` of incoming messages implements `Unpin`, thus there
is no need to use `pin_project` on the `UntilImported` struct. Instead
one only has to add the `Unpin` trait bound on the `I` trait parameter.
* client/network-gossip/bridge: Use bounded channel
Instead of returning an unbounded channel on
`GossipEngine::messages_for` return a bounded channel. For now the
channel length is determined by the amount of past messages cached in
the `ConsensusGossip`.
With a bounded channel, one can't just fire-and-forget style send into
it, but has to first check whether the channel is ready. Thus this
commit restructures `GossipEngine::poll` and introduces a
`ForwardingState` into `GossipEngine`.
* client/network-gossip/bridge: Add quickcheck for different size channels
Previously, when publishing ones address onto the DHT, the signature
signing those addresses would be SCALE encoded twice.
This commit removes the second encoding and adjusts the tests
to catch future regressions.
Instead of logging value-found-event-handling failures or value-put
failures on error level, log them on debug level only additionally
recording them via Prometheus.
Motivation is that both events can happen in "normal" operations and
thus clutter the logs.
* client/network-gossip/src/state_machine: Add unit test for on_incoming
Add two unit tests to ensure `on_incoming` is ingoring discarded
messages and reports and ignores messages from unknown peers.
* client/network-gossip/state_machine: Reduce indentation level
* client/network-gossip/bridge: Remove unnecessary into_iter
* client/network-gossip/state_machine: Report success after register check
* client/network-gossip/state_machine: Error not report on unregistered
`peers` contains all the peers we're connected to. If we receive a
message from a peer not in this list, that means there's an internal
problem within the local client. It's not the fault of the peer in
question.
With the above in mind instead of reducing the reputation of such peer,
log an error.
Expose the amount of addresses that are passed from the authority
discovery module down to the network peer set manager as a priority
group in order to connect to the corresponding nodes.
```
authority_discovery_priority_group_size 10
```
* client/network-gossip: Move sink IO outside of state_machine
`ConsensusGossip` is supposed to be a deterministic state machine.
`GossipEngine` wrapping `ConsensusGossip` should handle IO operations.
This commit moves the `message_sink` IO operations to `GossipEngine`.
More specifically on incoming messages a `GossipEngine` calls
`ConsensusGossip::on_incoming` to validate and register the messages.
`ConsensusGossip` returns the valid messages which are then forwarded by
`GossipEngine` to the upper layer via the `message_sinks`.
* client/network-gossip: Adjust and extend tests
* Update client/network-gossip/src/bridge.rs
Co-authored-by: Benjamin Kampmann <ben.kampmann@googlemail.com>
* service/src/lib: Enable authority discovery on sentry nodes
When run as a sentry node, the authority discovery module does not
publish any addresses to the dht, but still discovers validators and
sentry nodes of validators.
* bin/node/cli/src/service: Use expressions instead of statements
* Cargo.lock: Run `cargo update`
* service/src/lib: Fix compile error
* client/authority-discovery: Allow to be run by sentry node
When run as a sentry node, the authority discovery module does not
publish any addresses to the dht, but still discovers validators and
sentry nodes of validators.
* client/authority-discovery/src/lib: Wrap lines at 100 characters
* client/authority-discovery: Remove TODO and unused import
* client/authority-discovery: Pass role to new unit tests
* client/authority-discovery: Apply suggestions
Co-Authored-By: André Silva <123550+andresilva@users.noreply.github.com>
* bin/node/cli/src/service: Use expressions instead of statements
Co-authored-by: André Silva <123550+andresilva@users.noreply.github.com>
* client/authority-discovery: Rework error handling
Instead of `handle_dht_events` returning a `Result<(), Error>`, return
a `Poll<Error>` where `Poll::Pending` signals that there are no more
events to handle and `Poll::Ready(Error)` signals that a fatal error
occured. Non fatal errors are handled within `handle_dht_events`
directly, thus looping in `poll` is not necessary anymore.
* client/authority-discovery: Return () instead of error on termiantion
* Update client/authority-discovery/src/lib.rs
Co-Authored-By: André Silva <123550+andresilva@users.noreply.github.com>
Co-authored-by: Bastian Köcher <bkchr@users.noreply.github.com>
Co-authored-by: André Silva <123550+andresilva@users.noreply.github.com>
* Substrate Dashboard example
* Improve README
* Update README_dashboard.md
* Add screenshots
* Minor fix
* Minor fix, image link
* .maintain/sentry-node: Add monitoring to docker-compose stack
With this patch a user can run the following fully configured and
monitored setup with a single command:
`docker-compose -f .maintain/sentry-node/docker-compose.yml up`
- 2 validators in two different network namespaces, connected via one
sentry node.
- Polkadot-js/apps to connect to one of the nodes above.
- Prometheus scraping the 3 Substrate nodes.
- Grafana displaying data from Prometheus with community dashboards
* .maintain/monitoring/grafana: Change default datasource name
* .maintain/monitoring/grafana: Add metric namespace option
* .maintain/monitoring/grafana: Remove `host` metric from most metrics
* .maintain/monitoring/grafana: Remove underscore from metric_namespace
* .maintain/monitoring: Use `instance` label instead of `hostname`
To identify a scrape target, one should use `instance` and not
`hostname` as multiple targets might run on the same node.
See https://prometheus.io/docs/concepts/jobs_instances/ for details.
* .maintain/monitoring: Introduce instance variable
* .maintain/monitoring/grafana: Rename substrate_block_height_number
* .maitain/monitoring/grafana: Use instance instead of host in legend
* .maintain/monitoring: Remove node exporter dependency
* .maintain/sentry-node/prometheus: Simplify configuration
* .maintain/monitoring/grafana: Update README and remove images
* .maintain/sentry-node: Improve docs
* .maintain/monitoring/grafana: Use metric_namespace template variable
* Use --sentry from v0.7.29 instead of a reserved-node
* .maintain/sentry-node: Revert sentry-a using validator-b as bootnode
Co-authored-by: DerFredy - @derfredy:matrix.org <derfredy@gmail.com>
Co-authored-by: david <davidd@custom.home>
* client/authority-discovery: Add test ensure termination on network termi
* client/authority-discovery: Terminate when network does
When the dht event stream returns Poll::Ready(None) it is likely due to
the network terminating. When the network terminates due to the node
itself shutting down or due to a fatal error, there is no purpose in
continuing to run the authority discovery module.
* client/authority-discovery/src/lib: Apply suggestions
Co-Authored-By: André Silva <123550+andresilva@users.noreply.github.com>
Co-authored-by: André Silva <123550+andresilva@users.noreply.github.com>
* client/finality-grandpa: Instrument until-imported queue
The `UntilImported` queue takes as input finality grandpa messages that
depend on blocks that are not yet imported and holds them back until
those blocks are imported.
This commit adds a basic metric, the amount of messages waiting in the
queue, to the module. For now this metric is only available for the
global `UntilImported` queue awaiting blocks for commit and catch-up
messages.
* client/finality-grandpa/src/until_imported: Update metric help text
Co-Authored-By: Ashley <ashley.ruglys@gmail.com>
Co-authored-by: Ashley <ashley.ruglys@gmail.com>
The library `sysinfo` exposes process memory as kibibytes and not bytes,
thus the value needs to be multiplied by 1024 to comply with the metric
name and the Prometheus base units [1].
[1] https://prometheus.io/docs/practices/naming/#base-units
* client/finality-grandpa/src/until_imported: Refactor schedule_wait
Previously `BlockUntilImported::schedule_wait` took two closures, one to
report ready items and one to report items to await. None of the implementors of
`BlockUntilImported` call both closures. From a symantic perspective it
would as well not make sense to both await something and state that it
is ready.
Instead with this commit `BlockUntilImported::schedule_wait` simply
returns whether the given item needs waiting or is ready to be passed
on.
This reduces complexity by:
- Removing side effects through the two closures.
- Reducing borrowing given that `UntilImported` `ready` and `pending`
don't need to be borrowed from within the two closures.
- Removes the need for trait bounds for the two closures.
* client/finality-grandpa/src/until_imported: Fix comments
Co-Authored-By: André Silva <andre.beat@gmail.com>
Co-authored-by: André Silva <andre.beat@gmail.com>
* client/finality-grandpa/src/until_imported: Refactor BlockGlobalMessage
`BlockGlobalMessage` owns an `inner` which contains (1) a count for the
amount of outstanding blocks to be waited on and (2) the message itself.
Given that both is already wrapped in an `Arc` there is no need to keep
track of the outstanding blocks, given that it simply corresponds to the
amount of strong reference counts on the `Arc` itself.
This commit removes the atomic counter within `inner` and piggy backs on
the `Arc` reference counter instead.
* client/finality-grandpa/src/until_imported: Remove useless match
* client/finality-grandpa/src/until_imported.rs: Remove unused var
Co-Authored-By: André Silva <andre.beat@gmail.com>
* client/finality-grandpa/src/until_imported: Address comment suggestion
Co-authored-by: André Silva <andre.beat@gmail.com>
A gossip validator returns `ValidationResult::Discard` for messages that
should not be processed nor kept due to not being useful for the node.
Examples are out-of-scope messages or messages with bad signatures. The
corresponding log message did not reflect this, thus the update through
this patch.
One caveat is a finality grandpa neighbor packet. Given that the concept
of neighbor packets is handled in the finality grandpa gossip validator
only, valid neighbor packets are always discarded on the
`client/network-gossip` layer.
* client/network-gossip/src/bridge: Finish when network even stream closes
Previously within `<GossipEngine as Future>::poll` one would poll the
`network_event_stream` ignoring all messages other than
`Poll::Ready(Some())`. Ignoring `Poll::Ready(None)` leads to a panic on
the next poll of the stream, gien that it is not fused.
By design `network_event_stream` does not close unless an unbounded send
into it fails, or the `NetworkWorker` gets shut down.
> The stream never ends (unless the `NetworkWorker` gets shut down).
> (client/network/src/service.rs)
An `unbounded_send` to fail on an unbounded channel is unlikely. The
`NetworkWorker` shutting down is not unlikely. In such case the
`GossipEngine` should shut down as well.
With this patch a `<GossipEngine as Future>` finishes on
`Poll::Ready(None)` returned from `network_event_stream`.
* client/finality-grandpa/communication: Error on gossip engine finished
Have `<NetworkBridge as Future>::poll` return `Poll::Ready(Err)` instead
of `Poll::Ready(Ok)` to be consistent with the handling of the neighbor
packet worker stream and the gossip validator report stream. Both `Err`
as well as `Ok` shut down the `NetworkBridge` as well as the
`VoterWorker`.
* client/network-gossip/src/bridge: Add regression test
* client/network-gossip: Move substrate test client to dev dependencies
* client/network-gossip: Remove TODO
Addressed in a follow up pull request.
* client/network-gossip/bridge: Put match on newline after loop
* client/finality-grandpa/src/observer: Fix regression test
Make sure the event stream sender side is not dropped till the end.
* client/network/src/protocol: Start Prometheus metric help with capital
* client/network/src/protocol: Differentiate metric status as label
Prometheus query language is powerful through its multi-dimensional data
model. Metric names are hirarchical whereas labels enable data to become
multi-dimensional.
Exposing the justification of finality-proof status as a label allows
for more powerful queries.
* client/network/src/protocol: Remove 'Total' from non counter metric help
The word 'total' is reserved for accumulating counters. Counters have to
be monotonically increasing. `obsolete_requests` can decrease, thereby
it is defined as a `Gauge` and not a `Counter`.
For more details on metric naming see
https://prometheus.io/docs/practices/naming/
* client/finality-grandpa: Add Prometheus metrics to GossipValidator
Instrument finality grandpa `GossipValidator` exposing count of messages
validated by message type and message action.
```
\# HELP substrate_finality_grandpa_communication_gossip_validator_messages Number of messages validated by the finality grandpa gossip validator.
\# TYPE substrate_finality_grandpa_communication_gossip_validator_messages counter
substrate_finality_grandpa_communication_gossip_validator_messages{action="discard",message="neighbor"} 39
substrate_finality_grandpa_communication_gossip_validator_messages{action="keep",message="vote"} 28
```
* client/finality-grandpa: Add None as Prometheus registry in tests
* client/finality-granpda/src/communication: Refactor metric registration
* client/service/src/builder.rs: Add build_info metric
Add static Prometheus metric exposing the chain name, the version and
the commit.
* client/service/src/builder.rs: Move node_role to static metrics
The Prometheus metrics `node_role` is static and thus there is no need
to keep a reference of it within `ServiceMetrics`. This follows the
example of the `build_info` metric.
* client/service/src/builder.rs: Adjust indentation