Add docs for SSZ Query package (#16299)

**What type of PR is this?**

Documentation

**What does this PR do? Why is it needed?**

Although godoc and comments are well-written in `encoding/ssz/query`
package, we (@rkapka, @fernantho, @syjn99)
[agreed](https://discord.com/channels/476244492043812875/1387734369527136297/1466075406523174944)
that it would be great to have human-readable documentation.

**Which issues(s) does this PR fix?**

Part of  #15587 & #15598 

**Other notes for review**

This documentation is first drafted by Claude Code, and then has a few
rounds of self-review.

**Acknowledgements**

- [x] I have read
[CONTRIBUTING.md](https://github.com/prysmaticlabs/prysm/blob/develop/CONTRIBUTING.md).
- [x] I have included a uniquely named [changelog fragment
file](https://github.com/prysmaticlabs/prysm/blob/develop/CONTRIBUTING.md#maintaining-changelogmd).
- [x] I have added a description with sufficient context for reviewers
to understand this PR.
- [ ] I have tested that my changes work as expected and I added a
testing plan to the PR description (if applicable).

---------

Co-authored-by: fernantho <fernantho1@gmail.com>
Co-authored-by: Radosław Kapka <radoslaw.kapka@gmail.com>

api/client/event/event_stream.go

@@ -34,18 +34,6 @@ type Event struct {
 	Data      []byte
 }
 
-// PublishEvent enqueues an event without blocking the producer. If the channel is full,
-// the event is dropped since only the most recent heads are relevant.
-func PublishEvent(eventsChannel chan<- *Event, event *Event) {
-	if eventsChannel == nil || event == nil {
-		return
-	}
-	select {
-	case eventsChannel <- event:
-	default:
-	}
-}
-
 // EventStream is responsible for subscribing to the Beacon API events endpoint
 // and dispatching received events to subscribers.
 type EventStream struct {
@@ -79,20 +67,19 @@ func (h *EventStream) Subscribe(eventsChannel chan<- *Event) {
 	fullUrl := h.host + "/eth/v1/events?topics=" + allTopics
 	req, err := http.NewRequestWithContext(h.ctx, http.MethodGet, fullUrl, nil)
 	if err != nil {
-		PublishEvent(eventsChannel, &Event{
+		eventsChannel <- &Event{
 			EventType: EventConnectionError,
 			Data:      []byte(errors.Wrap(err, "failed to create HTTP request").Error()),
-		})
-		return
+		}
 	}
 	req.Header.Set("Accept", api.EventStreamMediaType)
 	req.Header.Set("Connection", api.KeepAlive)
 	resp, err := h.httpClient.Do(req)
 	if err != nil {
-		PublishEvent(eventsChannel, &Event{
+		eventsChannel <- &Event{
 			EventType: EventConnectionError,
 			Data:      []byte(errors.Wrap(err, client.ErrConnectionIssue.Error()).Error()),
-		})
+		}
 		return
 	}
 
@@ -113,6 +100,7 @@ func (h *EventStream) Subscribe(eventsChannel chan<- *Event) {
 		select {
 		case <-h.ctx.Done():
 			log.Info("Context canceled, stopping event stream")
+			close(eventsChannel)
 			return
 		default:
 			line := scanner.Text()
@@ -121,7 +109,7 @@ func (h *EventStream) Subscribe(eventsChannel chan<- *Event) {
 				// Empty line indicates the end of an event
 				if eventType != "" && data != "" {
 					// Process the event when both eventType and data are set
-					PublishEvent(eventsChannel, &Event{EventType: eventType, Data: []byte(data)})
+					eventsChannel <- &Event{EventType: eventType, Data: []byte(data)}
 				}
 
 				// Reset eventType and data for the next event
@@ -142,9 +130,9 @@ func (h *EventStream) Subscribe(eventsChannel chan<- *Event) {
 	}
 
 	if err := scanner.Err(); err != nil {
-		PublishEvent(eventsChannel, &Event{
+		eventsChannel <- &Event{
 			EventType: EventConnectionError,
 			Data:      []byte(errors.Wrap(err, errors.Wrap(client.ErrConnectionIssue, "scanner failed").Error()).Error()),
-		})
+		}
 	}
 }

api/client/event/event_stream_test.go

@@ -53,7 +53,7 @@ func TestEventStream(t *testing.T) {
 	defer server.Close()
 
 	topics := []string{"head"}
-	eventsChannel := make(chan *Event, 4)
+	eventsChannel := make(chan *Event, 1)
 	stream, err := NewEventStream(t.Context(), http.DefaultClient, server.URL, topics)
 	require.NoError(t, err)
 	go stream.Subscribe(eventsChannel)
@@ -80,7 +80,7 @@ func TestEventStream(t *testing.T) {
 
 func TestEventStreamRequestError(t *testing.T) {
 	topics := []string{"head"}
-	eventsChannel := make(chan *Event, 4)
+	eventsChannel := make(chan *Event, 1)
 	ctx := t.Context()
 
 	// use valid url that will result in failed request with nil body

changelog/jtraglia-add-specrefs-readme.md

@@ -0,0 +1,3 @@
+### Added
+
+- Added README for maintaining specrefs.

changelog/jtraglia-nightly-reftests-with-run-id.md

@@ -0,0 +1,3 @@
+### Added
+
+- The ability to download the nightly reference tests from a specific day.

changelog/syjn99_docs-ssz-ql.md

@@ -0,0 +1,3 @@
+### Ignored
+
+- Add handy documentation for SSZ Query package (`encoding/ssz/query`).

encoding/ssz/query/doc.md

@@ -0,0 +1,190 @@
+# SSZ Query Package
+
+The `encoding/ssz/query` package provides a system for analyzing and querying SSZ ([Simple Serialize](https://github.com/ethereum/consensus-specs/blob/master/ssz/simple-serialize.md)) data structures, as well as generating Merkle proofs from them. It enables runtime analysis of SSZ-serialized Go objects with reflection, path-based queries through nested structures, generalized index calculation, and Merkle proof generation.
+
+This package is designed to be generic. It operates on arbitrary SSZ-serialized Go values at runtime, so the same query/proof machinery applies equally to any SSZ type, including the BeaconState/BeaconBlock.
+
+## Usage Example
+
+```go
+// 1. Analyze an SSZ object
+block := &ethpb.BeaconBlock{...}
+info, err := query.AnalyzeObject(block)
+
+// 2. Parse a path
+path, err := query.ParsePath(".body.attestations[0].data.slot")
+
+// 3. Get the generalized index
+gindex, err := query.GetGeneralizedIndexFromPath(info, path)
+
+// 4. Generate a Merkle proof
+proof, err := info.Prove(gindex)
+
+// 5. Get offset and length to slice the SSZ-encoded bytes
+sszBytes, _ := block.MarshalSSZ()
+_, offset, length, err := query.CalculateOffsetAndLength(info, path)
+// slotBytes contains the SSZ-encoded value at the queried path
+slotBytes := sszBytes[offset : offset+length]
+```
+
+## Exported API
+
+The main exported API consists of:
+
+```go
+// AnalyzeObject analyzes an SSZ object and returns its structural information
+func AnalyzeObject(obj SSZObject) (*SszInfo, error)
+
+// ParsePath parses a path string like ".field1.field2[0].field3"
+func ParsePath(rawPath string) (Path, error)
+
+// CalculateOffsetAndLength computes byte offset and length for a path within an SSZ object
+func CalculateOffsetAndLength(sszInfo *SszInfo, path Path) (*SszInfo, uint64, uint64, error)
+
+// GetGeneralizedIndexFromPath calculates the generalized index for a given path
+func GetGeneralizedIndexFromPath(info *SszInfo, path Path) (uint64, error)
+
+// Prove generates a Merkle proof for a target generalized index
+func (s *SszInfo) Prove(gindex uint64) (*fastssz.Proof, error)
+```
+
+## Type System
+
+### SSZ Types
+
+The package now supports [all standard SSZ types](https://github.com/ethereum/consensus-specs/blob/master/ssz/simple-serialize.md#typing) except `ProgressiveList`, `ProgressiveContainer`, `ProgressiveBitlist`, `Union`, and `CompatibleUnion`.
+
+### Core Data Structures
+
+#### `SszInfo`
+
+The `SszInfo` structure contains complete structural metadata for an SSZ type:
+
+```go
+type SszInfo struct {
+   sszType       SszType           // SSZ Type classification
+   typ           reflect.Type      // Go reflect.Type
+   source        SSZObject         // Original SSZObject reference. Mostly used for reusing SSZ methods like `HashTreeRoot`.
+   isVariable    bool              // True if contains variable-size fields
+
+   // Composite types have corresponding metadata. Other fields would be nil except for the current type.
+   containerInfo *containerInfo
+   listInfo      *listInfo
+   vectorInfo    *vectorInfo
+   bitlistInfo   *bitlistInfo
+   bitvectorInfo *bitvectorInfo
+}
+```
+
+#### `Path`
+
+The `Path` structure represents navigation paths through SSZ structures. It supports accessing a field by field name, accessing an element by index (list/vector type), and finding the length of homogenous collection types. The `ParsePath` function parses a raw string into a `Path` instance, which is commonly used in other APIs like `CalculateOffsetAndLength` and `GetGeneralizedIndexFromPath`.
+
+```go
+type Path struct {
+   Length   bool           // Flag for length queries (e.g., len(.field))
+   Elements []PathElement  // Sequence of field accesses and indices
+}
+
+type PathElement struct {
+   Name  string  // Field name
+   Index *uint64 // list/vector index (nil if not an index access)
+}
+```
+
+## Implementation Details
+
+### Type Analysis (`analyzer.go`)
+
+The `AnalyzeObject` function performs recursive type introspection using Go reflection:
+
+1. **Type Inspection** - Examines Go `reflect.Value` to determine SSZ type
+   - Basic types (`uint8`, `uint16`, `uint32`, `uint64`, `bool`): `SSZType` constants
+   - Slices: Determined from struct tags (`ssz-size` for vectors, `ssz-max` for lists). There is a related [write-up](https://hackmd.io/@junsong/H101DKnwxl) regarding struct tags.
+   - Structs: Analyzed as Containers with field ordering from JSON tags
+   - Pointers: Dereferenced automatically
+
+2. **Variable-Length Population** - Determines actual sizes at runtime
+   - For lists: Iterates elements, caches sizes for variable-element lists
+   - For containers: Recursively populates variable fields, adjusts offsets
+   - For bitlists: Decodes bit length from bitvector
+
+3. **Offset Calculation** - Computes byte positions within serialized data
+   - Fixed-size fields: Offset = sum of preceding field sizes
+   - Variable-size fields: Offset stored as 4-byte pointer entries
+
+### Path Parsing (`path.go`)
+
+The `ParsePath` function parses path strings with the following rules:
+
+- **Dot notation**: `.field1.field2` for field access
+- **Array indexing**: `[0]`, `[42]` for element access
+- **Length queries**: `len(.field)` for list/vector lengths
+- **Character set**: Only `[A-Za-z0-9._\[\]\(\)]` allowed
+
+Example:
+```go
+path, _ := ParsePath(".nested.array_field[5].inner_field")
+// Returns: Path{
+//   Elements: [
+//     PathElement{Name: "nested"},
+//     PathElement{Name: "array_field", Index: <Pointer to uint64(5)>},
+//     PathElement{Name: "inner_field"}
+//   ]
+// }
+```
+
+### Generalized Index Calculation (`generalized_index.go`)
+
+The generalized index is a tree position identifier. This package follows the [Ethereum consensus-specs](https://github.com/ethereum/consensus-specs/blob/master/ssz/merkle-proofs.md#generalized-merkle-tree-index) to calculate the generalized index.
+
+### Merkle Proof Generation (`merkle_proof.go`, `proof_collector.go`)
+
+The `Prove` method generates Merkle proofs using a single-sweep merkleization algorithm:
+
+#### Algorithm Overview
+
+**Key Terms:**
+
+- **Target gindex** (generalized index): The position of the SSZ element you want to prove, expressed as a generalized Merkle tree index. Stored in `Proof.Index`.
+  - Note: The generalized index for root is 1.
+- **Registered gindices**: The set of tree positions whose node hashes must be captured during merkleization in order to later assemble the proof.
+- **Sibling node**: The node that shares the same parent as another node.
+- **Leaf value**: The 32-byte hash of the target node (the node being proven). Stored in `Proof.Leaf`.
+
+**Phases:**
+
+1. **Registration Phase** (`addTarget`)
+> Goal: determine exactly which sibling hashes are needed for the proof.
+
+   - Record the target gindex as the proof target.
+   - Starting from the target node, walk the Merkle tree from the leaf (target gindex) to the root (gindex = 1).
+   - At each step:
+     - Compute and register the sibling gindex (`i XOR 1`) as “must collect”.
+     - Move to the parent (`i = i/2`).
+   - This produces the full set of registered gindices (the sibling nodes on the target-to-root path).
+
+2. **Merkleization Phase** (`merkleize`)
+> Goal: recursively merkleize the tree and capture the needed hashes.
+
+   - Recursively traverse the SSZ structure and compute Merkle tree node hashes from leaves to root.
+   - Whenever the traversal computes a node whose gindex is in registered gindices, store that node’s hash for later proof construction.
+
+3. **Proof Assembly Phase** (`toProof`)
+> Goal: create the final `fastssz.Proof` object in the correct format and order.
+
+```go
+// Proof represents a merkle proof against a general index.
+type Proof struct {
+	Index  int
+	Leaf   []byte
+	Hashes [][]byte
+}
+```
+
+   - Set `Proof.Index` to the target gindex.
+   - Set `Proof.Leaf` to the 32-byte hash of the target node.
+   - Build `Proof.Hashes` by walking from the target node up to (but not including) the root:
+     - At node `i`, append the stored hash for the sibling (`i XOR 1`).
+     - Move to the parent (`i = i/2`).
+   - The resulting `Proof.Hashes` is ordered from the target level upward, containing one sibling hash per tree level on the path to the root.

specrefs/README.md

@@ -0,0 +1,35 @@
+# Specification References
+
+This directory contains specification reference tracking files managed by
+[ethspecify](https://github.com/jtraglia/ethspecify).
+
+## Installation
+
+Install `ethspecify` with the following command:
+
+```bash
+pipx install ethspecify
+```
+
+> [!NOTE]
+> You can run `ethspecify <cmd>` in the `specrefs` directory or
+> `ethspecify <cmd> --path=specrefs` from the project's root directory.
+
+## Maintenance
+
+When adding support for a new specification version, follow these steps:
+
+0. Change directory into the `specrefs` directory.
+1. Update the version in `.ethspecify.yml` configuration.
+2. Run `ethspecify process` to update/populate specrefs.
+3. Run `ethspecify check` to check specrefs.
+4. If there are errors, use the error message as a guide to fix the issue. If
+   there are new specrefs with empty sources, implement/locate each item and
+   update each specref source list. If you choose not to implement an item,
+   add an exception to the appropriate section the the `.ethspecify.yml`
+   configuration.
+5. Repeat steps 3 and 4 until `ethspecify check` passes.
+6. Run `git diff` to view updated specrefs. If an object/function/etc has
+   changed, make the necessary updates to the implementation.
+7. Lastly, in the project's root directory, run `act -j check-specrefs` to
+   ensure everything is correct.

testing/spectest/README.md

@@ -21,10 +21,14 @@ There are tests for mainnet and minimal config, so for each config we will add a
 
 ## Running nightly spectests
 
-Since [PR 15312](https://github.com/OffchainLabs/prysm/pull/15312), Prysm has support to download "nightly" spectests from github via a starlark rule configuration by environment variable. 
-Set `--repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly` when running spectest to download the "nightly" spectests. 
-Note: A GITHUB_TOKEN environment variable is required to be set. The github token must be a [fine grained token](https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/managing-your-personal-access-tokens#creating-a-fine-grained-personal-access-token).
+Since [PR 15312](https://github.com/OffchainLabs/prysm/pull/15312), Prysm has support to download "nightly" spectests from github via a starlark rule configuration by environment variable.
+Set `--repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly` or `--repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly-<run_id>` when running spectest to download the "nightly" spectests.
+Note: A GITHUB_TOKEN environment variable is required to be set. The github token does not need to be associated with your main account; it can be from a "burner account". And the token does not need to be a fine-grained token; it can be a classic token.
 
 ```
 bazel test //... --test_tag_filters=spectest --repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly
 ```
+
+```
+bazel test //... --test_tag_filters=spectest --repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly-21422848633
+```

tools/download_spectests.bzl

@@ -1,5 +1,6 @@
 # bazel build @consensus_spec_tests//:test_data
 # bazel build @consensus_spec_tests//:test_data --repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly
+# bazel build @consensus_spec_tests//:test_data --repo_env=CONSENSUS_SPEC_TESTS_VERSION=nightly-<run_id>
 
 def _get_redirected_url(repository_ctx, url, headers):
     if not repository_ctx.which("curl"):
@@ -24,7 +25,7 @@ def _impl(repository_ctx):
     version = repository_ctx.getenv("CONSENSUS_SPEC_TESTS_VERSION") or repository_ctx.attr.version
     token = repository_ctx.getenv("GITHUB_TOKEN") or ""
 
-    if version == "nightly":
+    if version == "nightly" or version.startswith("nightly-"):
         print("Downloading nightly tests")
         if not token:
             fail("Error GITHUB_TOKEN is not set")
@@ -34,16 +35,22 @@ def _impl(repository_ctx):
             "Accept": "application/vnd.github+json",
         }
 
-        repository_ctx.download(
-            "https://api.github.com/repos/%s/actions/workflows/%s/runs?branch=%s&status=success&per_page=1"
-                % (repository_ctx.attr.repo, repository_ctx.attr.workflow, repository_ctx.attr.branch),
-            headers = headers,
-            output = "runs.json"
-        )
+        if version.startswith("nightly-"):
+            run_id = version.split("nightly-", 1)[1]
+            if not run_id:
+                fail("Error invalid run id")
+        else:
+            repository_ctx.download(
+                "https://api.github.com/repos/%s/actions/workflows/%s/runs?branch=%s&status=success&per_page=1"
+                    % (repository_ctx.attr.repo, repository_ctx.attr.workflow, repository_ctx.attr.branch),
+                headers = headers,
+                output = "runs.json"
+            )
 
-        run_id = json.decode(repository_ctx.read("runs.json"))["workflow_runs"][0]["id"]
-        repository_ctx.delete("runs.json")
+            run_id = json.decode(repository_ctx.read("runs.json"))["workflow_runs"][0]["id"]
+            repository_ctx.delete("runs.json")
 
+        print("Run id:", run_id)
         repository_ctx.download(
             "https://api.github.com/repos/%s/actions/runs/%s/artifacts"
                 % (repository_ctx.attr.repo, run_id),
@@ -108,8 +115,8 @@ consensus_spec_tests = repository_rule(
         "version": attr.string(mandatory = True),
         "flavors": attr.string_dict(mandatory = True),
         "repo": attr.string(default = "ethereum/consensus-specs"),
-        "workflow": attr.string(default = "generate_vectors.yml"),
-        "branch": attr.string(default = "dev"),
+        "workflow": attr.string(default = "nightly-reftests.yml"),
+        "branch": attr.string(default = "master"),
         "release_url_template": attr.string(default = "https://github.com/ethereum/consensus-specs/releases/download/%s"),
     },
 )

validator/client/grpc-api/grpc_validator_client.go

@@ -285,10 +285,10 @@ func (c *grpcValidatorClient) StartEventStream(ctx context.Context, topics []str
 	ctx, span := trace.StartSpan(ctx, "validator.gRPCClient.StartEventStream")
 	defer span.End()
 	if len(topics) == 0 {
-		eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+		eventsChannel <- &eventClient.Event{
 			EventType: eventClient.EventError,
 			Data:      []byte(errors.New("no topics were added").Error()),
-		})
+		}
 		return
 	}
 	// TODO(13563): ONLY WORKS WITH HEAD TOPIC.
@@ -299,10 +299,10 @@ func (c *grpcValidatorClient) StartEventStream(ctx context.Context, topics []str
 		}
 	}
 	if !containsHead {
-		eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+		eventsChannel <- &eventClient.Event{
 			EventType: eventClient.EventConnectionError,
 			Data:      []byte(errors.Wrap(client.ErrConnectionIssue, "gRPC only supports the head topic, and head topic was not passed").Error()),
-		})
+		}
 	}
 	if containsHead && len(topics) > 1 {
 		log.Warn("gRPC only supports the head topic, other topics will be ignored")
@@ -310,10 +310,10 @@ func (c *grpcValidatorClient) StartEventStream(ctx context.Context, topics []str
 
 	stream, err := c.beaconNodeValidatorClient.StreamSlots(ctx, &ethpb.StreamSlotsRequest{VerifiedOnly: true})
 	if err != nil {
-		eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+		eventsChannel <- &eventClient.Event{
 			EventType: eventClient.EventConnectionError,
 			Data:      []byte(errors.Wrap(client.ErrConnectionIssue, err.Error()).Error()),
-		})
+		}
 		return
 	}
 	c.isEventStreamRunning = true
@@ -327,25 +327,25 @@ func (c *grpcValidatorClient) StartEventStream(ctx context.Context, topics []str
 			if ctx.Err() != nil {
 				c.isEventStreamRunning = false
 				if errors.Is(ctx.Err(), context.Canceled) {
-					eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+					eventsChannel <- &eventClient.Event{
 						EventType: eventClient.EventConnectionError,
 						Data:      []byte(errors.Wrap(client.ErrConnectionIssue, ctx.Err().Error()).Error()),
-					})
+					}
 					return
 				}
-				eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+				eventsChannel <- &eventClient.Event{
 					EventType: eventClient.EventError,
 					Data:      []byte(ctx.Err().Error()),
-				})
+				}
 				return
 			}
 			res, err := stream.Recv()
 			if err != nil {
 				c.isEventStreamRunning = false
-				eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+				eventsChannel <- &eventClient.Event{
 					EventType: eventClient.EventConnectionError,
 					Data:      []byte(errors.Wrap(client.ErrConnectionIssue, err.Error()).Error()),
-				})
+				}
 				return
 			}
 			if res == nil {
@@ -357,15 +357,15 @@ func (c *grpcValidatorClient) StartEventStream(ctx context.Context, topics []str
 				CurrentDutyDependentRoot:  hexutil.Encode(res.CurrentDutyDependentRoot),
 			})
 			if err != nil {
-				eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+				eventsChannel <- &eventClient.Event{
 					EventType: eventClient.EventError,
 					Data:      []byte(errors.Wrap(err, "failed to marshal Head Event").Error()),
-				})
+				}
 			}
-			eventClient.PublishEvent(eventsChannel, &eventClient.Event{
+			eventsChannel <- &eventClient.Event{
 				EventType: eventClient.EventHead,
 				Data:      b,
-			})
+			}
 		}
 	}
 }

validator/client/grpc-api/grpc_validator_client_test.go

@@ -223,7 +223,7 @@ func TestStartEventStream(t *testing.T) {
 
 	for _, tc := range tests {
 		t.Run(tc.name, func(t *testing.T) {
-			eventsChannel := make(chan *eventClient.Event, 4) // Buffer to prevent blocking
+			eventsChannel := make(chan *eventClient.Event, 1) // Buffer to prevent blocking
 			tc.prepare()                                      // Setup mock expectations
 
 			go grpcClient.StartEventStream(ctx, tc.topics, eventsChannel)

validator/client/runner_test.go

@@ -441,6 +441,7 @@ func TestRunnerPushesProposerSettings_ValidContext(t *testing.T) {
 		defer assertValidContext(t, timedCtx, ctx)
 		delay(t)
 	})
+	vcm.EXPECT().EventStreamIsRunning().Return(true).AnyTimes().Do(func() { delay(t) })
 	vcm.EXPECT().SubmitValidatorRegistrations(liveCtx, gomock.Any()).Do(func(ctx context.Context, _ any) {
 		defer assertValidContext(t, timedCtx, ctx) // This is the specific regression test assertion for PR 15369.
 		delay(t)

validator/client/service.go

@@ -66,8 +66,6 @@ type ValidatorService struct {
 	closeClientFunc         func() // validator client stop function is used here
 }
 
-const eventChannelBufferSize = 32
-
 // Config for the validator service.
 type Config struct {
 	Validator               iface.Validator
@@ -236,7 +234,7 @@ func (v *ValidatorService) Start() {
 		distributed:                    v.distributed,
 		disableDutiesPolling:           v.disableDutiesPolling,
 		accountsChangedChannel:         make(chan [][fieldparams.BLSPubkeyLength]byte, 1),
-		eventsChannel:                  make(chan *eventClient.Event, eventChannelBufferSize),
+		eventsChannel:                  make(chan *eventClient.Event, 1),
 	}
 
 	hm := newHealthMonitor(v.ctx, v.cancel, v.maxHealthChecks, v.validator)

validator/client/validator.go

@@ -64,11 +64,6 @@ var (
 	msgNoKeysFetched     = "No validating keys fetched. Waiting for keys..."
 )
 
-const (
-	eventStreamStopped uint32 = iota
-	eventStreamRunning
-)
-
 type validator struct {
 	logValidatorPerformance            bool
 	distributed                        bool
@@ -87,7 +82,6 @@ type validator struct {
 	cachedAttestationData              *ethpb.AttestationData
 	accountsChangedChannel             chan [][fieldparams.BLSPubkeyLength]byte
 	eventsChannel                      chan *eventClient.Event
-	eventStreamState                   atomic.Uint32
 	highestValidSlot                   primitives.Slot
 	submittedAggregates                map[submittedAttKey]*submittedAtt
 	graffitiStruct                     *graffiti.Graffiti
@@ -1217,40 +1211,12 @@ func (v *validator) PushProposerSettings(ctx context.Context, slot primitives.Sl
 }
 
 func (v *validator) StartEventStream(ctx context.Context, topics []string) {
-	if !v.eventStreamState.CompareAndSwap(eventStreamStopped, eventStreamRunning) {
+	if v.EventStreamIsRunning() {
 		log.Debug("EventStream is already running")
 		return
 	}
 	log.WithField("topics", topics).Info("Starting event stream")
-	go v.runEventStream(ctx, topics)
-}
-
-func (v *validator) runEventStream(ctx context.Context, topics []string) {
-	defer v.eventStreamState.Store(eventStreamStopped)
-	backoff := time.Second
-	const maxBackoff = 30 * time.Second
-
-	for {
-		v.validatorClient.StartEventStream(ctx, topics, v.eventsChannel)
-		if ctx.Err() != nil {
-			return
-		}
-
-		log.WithField("retryIn", backoff).Warn("Event stream ended unexpectedly, attempting to resubscribe")
-		timer := time.NewTimer(backoff)
-		select {
-		case <-ctx.Done():
-			timer.Stop()
-			return
-		case <-timer.C:
-		}
-		if backoff < maxBackoff {
-			backoff *= 2
-			if backoff > maxBackoff {
-				backoff = maxBackoff
-			}
-		}
-	}
+	v.validatorClient.StartEventStream(ctx, topics, v.eventsChannel)
 }
 
 func (v *validator) checkDependentRoots(ctx context.Context, head *structs.HeadEvent) error {
@@ -1337,7 +1303,7 @@ func (v *validator) ProcessEvent(ctx context.Context, event *eventClient.Event)
 }
 
 func (v *validator) EventStreamIsRunning() bool {
-	return v.eventStreamState.Load() == eventStreamRunning
+	return v.validatorClient.EventStreamIsRunning()
 }
 
 func (v *validator) Host() string {