[kbhargava-jump]

No description provided.

[Copilot]

🟡 Changes recommended

Event pacing can crash or indefinitely throttle producers, and telemetry schema compatibility and vote-write accuracy need correction.

4 open findings

• Validate maximum configured rate to prevent zero nanoseconds per byte · New
• Preserve protobuf enum values when removing BOOTUP · New
• Avoid repeated producer throttling after ring-buffer timeout · New
• Record effective vote state values after updates · New

What changed in this PR

Expands runtime event reporting, improves bundle telemetry parity, and adds configurable event upload pacing.

Changes:

• Adds Alpenglow vote-write and richer reward/stake events.
• Captures bundle transaction diffs and stake mutations.
• Adds event backpressure, draining, pacing configuration, and tests.

File	Description
src/​flamenco/​stakes/​fd_stakes.c	Refines epoch vote counts and reward state initialization.
src/​flamenco/​runtime/​tests/​test_inflation_rewards.c	Tests zero-commission reward behavior.
src/​flamenco/​runtime/​tests/​test_accdb_svm.c	Tests zero-value vote credits.
src/​flamenco/​runtime/​test_bundle_exec.c	Tests bundle event parity.
src/​flamenco/​runtime/​Local.mk	Links bundle tests with Tango.
src/​flamenco/​runtime/​fd_runtime.h	Adds bundle event capture context.
src/​flamenco/​runtime/​fd_runtime_stack.h	Tracks commission entries.
src/​flamenco/​runtime/​fd_accdb_svm.c	Emits zero-value vote reward events.
src/​flamenco/​rewards/​fd_rewards.c	Handles zero-commission entries and delegation metadata.
src/​flamenco/​events/​fd_event_runtime.h	Declares new capture and vote-write APIs.
src/​flamenco/​events/​fd_event_runtime.c	Implements bundle, reward, and vote-write events.
src/​flamenco/​alpenglow/​fd_alpenglow.c	Reports certificate-driven vote writes.
src/​discof/​replay/​fd_replay_tile.c	Removes boot delegation baseline emission.
src/​discof/​execle/​fd_execle_tile.c	Captures bundle events.
src/​discof/​backtest/​fd_backtest_tile.c	Drains telemetry before exit.
src/​disco/​topo/​fd_topo.h	Adds event upload-rate topology state.
src/​disco/​events/​test_event_client.c	Updates pacing tests and constructor calls.
src/​disco/​events/​schema/​runtime_vote_write.json	Defines vote-write events.
src/​disco/​events/​schema/​runtime_txn.json	Clarifies transaction diff semantics.
src/​disco/​events/​schema/​runtime_stake_delegation.json	Removes boot baseline semantics.
src/​disco/​events/​schema/​runtime_reward.json	Adds delegated vote account metadata.
src/​disco/​events/​schema/​runtime_epoch.json	Updates VAT documentation.
src/​disco/​events/​schema/​runtime_block.json	References vote-write exclusions.
src/​disco/​events/​schema/​events.proto	Updates protobuf event contracts.
src/​disco/​events/​generated/​fd_event_gen.h	Updates generated event types.
src/​disco/​events/​generated/​fd_event_gen.c	Serializes new event fields.
src/​disco/​events/​generated/​fd_event_gen_test.h	Adds generated serialization coverage.
src/​disco/​events/​fd_event_tile.c	Returns event-link credit and configures pacing.
src/​disco/​events/​fd_event_report.h	Adds producer backpressure.
src/​disco/​events/​fd_event_report.c	Connects reporters to consumer progress.
src/​disco/​events/​fd_event_client.h	Exposes configurable upload rate.
src/​disco/​events/​fd_event_client.c	Implements pacing and batch draining.
src/​app/​shared/​fd_config.h	Adds upload-rate configuration.
src/​app/​shared/​fd_config_parse.c	Parses upload rate.
src/​app/​shared/​fd_config_json.c	Exports upload rate.
src/​app/​firedancer/​topology.c	Propagates upload rate.
src/​app/​firedancer/​config/​default.toml	Documents the new setting.
contrib/​blockstore/​blockstore2shredcap.c	Reports bank-hash sources and omissions.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

🔵 Needs a closer look

Event draining can lose unacknowledged data, pacing accepts crashing rates, and an existing protobuf enum value is renumbered.

3 open findings

• Preserve protobuf enum values when removing BOOTUP
• Validate maximum configured rate to prevent zero nanoseconds per byte
• Record effective vote state values after updates

1 resolved since last review

• Avoid repeated producer throttling after ring-buffer timeout

Previously missed (1)

In code that hasn't changed since last review

Drain waits only for unsent events, dropping unacknowledged events

src/​discof/​backtest/​fd_backtest_tile.c:371

This drain condition ignores events that have been sent but not yet acknowledged. QUEUE_UNSENT_COUNT reaches zero as soon as the circq cursor advances, while QUEUE_USED explicitly includes sent-but-unacknowledged events; the immediate process exit can therefore drop in-flight runtime events. Wait for QUEUE_USED to reach zero and report that pending count in timeout/disconnect warnings.

🧠 Review effort: Balanced

[Copilot]

🔵 Needs a closer look

Account-closure diffs and failed-redeem commission tracking can produce incorrect runtime events.

3 open findings

• Preserve protobuf enum values when removing BOOTUP
• Validate maximum configured rate to prevent zero nanoseconds per byte
• Record effective vote state values after updates

Previously missed (2)

In code that hasn't changed since last review

Normalize captured account post-state for closure events

src/​flamenco/​events/​fd_event_runtime.c:277

Bundle capture runs before commit. When a transaction drains an account to zero, it captures the old owner, data length and executable flag, but follower events run after fd_runtime_lthash_account() clears those fields. This makes leader and follower diffs disagree for account closures. Normalize the captured post-state without modifying the shared account.

Set commission flag only after successful reward redemption

src/​flamenco/​rewards/​fd_rewards.c:1446

When redeem_rewards() fails, the rent-adjustment or block-reward fallback can still reach this line with zero voter rewards. Those stakes need a partition entry, but did not produce a successful commission redeem. Setting this flag makes the new payout filters treat them as commission entries and can emit an extra zero vote-reward row. Set the flag only for successful redeems, preserving any entry established by another delegation.

🧠 Review effort: Balanced

[github-actions]

 ┌─ ⚡ PERF · 9ccca11 vs main@a480289 ─────────────────────────────────
 │ SUITE                               BASELINE          NEW         Δ
 │ replay tps, mainnet               14,459 tps   14,369 tps  ·  -0.62%
 │ bench tps, localnet              660,808 tps  656,103 tps  ·  -0.71%
 │ snapshot load, testnet                4.97 s       4.96 s  ·  -0.19%
 │ snapshot load, mainnet               67.57 s      68.28 s  ·  +1.05%
 │ epoch boundary, mainnet             144.6 ms     146.6 ms  ·  +1.36%
+│ mem total, mainnet                142.88 GiB   142.88 GiB  ▼   0.00%
+│ mem total, testnet                 80.19 GiB    80.19 GiB  ▼   0.00%
 │ clean compile, firedancer             3.26 s       3.25 s  ·  -0.23%
 │ binary size, firedancer             59.18 MB     59.21 MB  ·  +0.05%
 ├─────────────────────────────────────────────────────────────────────
@@ 0 REGRESSIONS · 0 WARNINGS · 2 IMPROVED · 7 NOISE @@
 └─────────────────────────────────────────────────────────────────────

history · 10 pushes

 ┌─ HISTORY · Δ vs main, per push, newest first ───────────────────────────────────────────
 │ HEAD         TPS    BENCH   SNAP·T   SNAP·M    EPOCH    MEM·M    MEM·T  COMPILE   BINARY
+│ 9ccca11   -0.62%   -0.71%   -0.19%   +1.05%   +1.36%    0.00%    0.00%   -0.23%   +0.05%
!│ 9806cbc   +0.84%   -0.58%   +0.01%   +0.10%   +1.23%    0.00%    0.00%   +0.02%   +0.05%
!│ 95bc2b7   -0.73%   -0.43%   -0.39%   +0.40%   +0.56%    0.00%    0.00%   -0.46%   +0.05%
+│ 1fbfedc   -0.14%   -0.44%   -0.37%        …        …    0.00%    0.00%   +0.86%   +0.05%
-│ 05357f0   +0.34%   +0.16%   +5.58%        …        …    0.00%    0.00%   +0.49%   +0.06%
-│ 707b371   +0.22%   -0.05%   +5.61%        …        …   +0.03%   +0.04%        …   +0.07%
!│ 929a4b7   -1.56%   +0.35%   +1.70%        …        …   +0.03%   +0.11%   +0.03%   +0.07%
!│ a71923d   +0.49%   +3.58%   +2.44%        …        …   +0.03%   +0.11%   +0.96%   +0.07%
!│ 1827508   +0.69%   -0.52%   +0.72%        …        …    0.00%    0.00%   +2.97%   +0.05%
!│ a4aada0   -0.55%   -0.70%   +2.67%        …        …    0.00%    0.00%   +1.62%   +0.05%
 └─────────────────────────────────────────────────────────────────────────────────────────

[Copilot]

🟡 Changes recommended

Protobuf enum renumbering breaks compatibility, and bundle account-closure events disagree with follower events.

5 open findings

• Preserve bootup enum value to maintain protobuf compatibility · New
• Preserve protobuf enum values when removing BOOTUP
• Validate maximum configured rate to prevent zero nanoseconds per byte
• Normalize closed account state in pre-commit capture · New
• Record effective vote state values after updates

🧠 Review effort: Balanced

[Copilot]

🔵 Needs a closer look

Upload-rate safety, protocol compatibility, and event post-state accuracy remain unresolved.

5 open findings

• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Validate maximum configured rate to prevent zero nanoseconds per byte
• Normalize closed account state in pre-commit capture
• Record effective vote state values after updates

🧠 Review effort: Balanced

[Copilot]

🔵 Needs a closer look

Reported compatibility and telemetry correctness concerns remain unresolved; their exclusion as duplicate feedback does not establish that they were fixed.

5 open findings

• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Validate maximum configured rate to prevent zero nanoseconds per byte
• Normalize closed account state in pre-commit capture
• Record effective vote state values after updates

🧠 Review effort: Balanced

[Copilot]

🟡 Changes recommended

Timestamp overflow and event-drain synchronization remain unresolved correctness issues.

4 open findings

• Prevent signed overflow when converting vote timestamps · New
• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Record effective vote state values after updates

2 resolved since last review

• Validate maximum configured rate to prevent zero nanoseconds per byte
• Normalize closed account state in pre-commit capture

Previously missed (1)

In code that hasn't changed since last review

Use a post-metrics completion marker for queue draining

src/​discof/​backtest/​fd_backtest_tile.c:374

A newer heartbeat does not guarantee that QUEUE_USED is current: src/disco/stem/fd_stem_run1.c:247–252 publishes the heartbeat before calling metrics_write, which updates the queue gauge. If events were consumed after the previous metrics update reported an empty queue, this loop can observe the new heartbeat with the old zero gauge and exit with events still pending. Use an explicit drain acknowledgement or a completion marker published after the queue metrics, rather than the generic heartbeat.

🧠 Review effort: Balanced

[Copilot]

🟡 Changes recommended

Event wire compatibility and drain-completion synchronization remain unresolved.

5 open findings

• Prevent signed overflow when converting vote timestamps
• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Wait for post-metrics heartbeat before exiting stem · New
• Record effective vote state values after updates

🧠 Review effort: Balanced

[Copilot]

🔵 Needs a closer look

The reported wire-compatibility and timestamp-overflow concerns remain unresolved; duplicate exclusion does not establish that they were fixed.

4 open findings

• Prevent signed overflow when converting vote timestamps
• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Record effective vote state values after updates

🧠 Review effort: Balanced

[nishk-jump]

add comment for 32 Mb/s (also makes it more obvious it bytes and not bits)

[Copilot]

🟡 Changes recommended

Migration-epoch vote-write events report older Tower credits instead of the updated Alpenglow credit total.

1 open issue

• Return the latest credits entry for duplicate migration epochs · New

4 resolved since last review

• Prevent signed overflow when converting vote timestamps
• Preserve bootup enum value to maintain protobuf compatibility
• Preserve protobuf enum values when removing BOOTUP
• Record effective vote state values after updates

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Previously reported wire-compatibility and timestamp-correctness issues remain unresolved in the supplied changes.

No open issues 🎉

1 resolved since last review

• Return the latest credits entry for duplicate migration epochs

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Unresolved critical issues break protobuf wire compatibility, introduce signed overflow for valid vote timestamps, and cause V4/V5 parity tests to fail.

2 open issues

• Emit vote timestamps directly in seconds · New
• Preserve identical update input for V4 and V5 parity · New

_{Review effort: Balanced}

[Copilot]

🔵 Needs a closer look

Removing bootup renumbers reward from wire value 4 to 3, breaking compatibility by causing collectors to misinterpret new and historical events.

2 open issues

• Preserve identical update input for V4 and V5 parity
• Emit vote timestamps directly in seconds

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Critical issues remain: removing bootup breaks protobuf enum compatibility, and closed delegated accounts can leave stale delegations in the event stream.

1 open issue

• Capture removals for previously delegated accounts · New

2 resolved since last review

• Preserve identical update input for V4 and V5 parity
• Emit vote timestamps directly in seconds

_{Review effort: Balanced}

[Copilot]

🟡 Changes recommended

Removing bootup renumbers reward from wire value 4 to 3, breaking compatibility with existing collectors and historical data.

1 open issue

• Skip credit history scans when runtime-diff reporting is disabled · New

1 resolved since last review

• Capture removals for previously delegated accounts

_{Review effort: Balanced}