Compare commits

..

11 Commits
3.4.3 ... flow

Author SHA1 Message Date
Alexey
f1bf95a7de Merge pull request #718 from astronaut808/fix/me-downstream-retry
Improve ME downstream retries for queued fairness backlog
2026-04-18 14:03:37 +03:00
Alexey
959a16af88 Merge pull request #716 from zarv1k/feature/configurable-proxy-confi-urls
feat: make URLS to obtain proxy_secret, getProxyConfig, getProxyConfgV6 files optionally configurable
2026-04-18 11:17:37 +03:00
Alexey
a54f9ba719 Merge branch 'flow' into feature/configurable-proxy-confi-urls 2026-04-18 11:16:38 +03:00
astronaut808
2d5cd9c8e1 Improve ME downstream retries for queued fairness backlog 2026-04-18 02:40:32 +05:00
Alexey
37b6f7b985 Weighted Fairness + 3-Leveled Pressure Model
Co-Authored-By: brekotis <93345790+brekotis@users.noreply.github.com>
2026-04-18 00:37:04 +03:00
Alexey
50e9e5cf32 Active Ring and DRR Hardening
Co-Authored-By: brekotis <93345790+brekotis@users.noreply.github.com>
2026-04-18 00:34:35 +03:00
Alexey
d72cfd6bc4 Merge branch 'flow' into feature/configurable-proxy-confi-urls 2026-04-17 19:44:46 +03:00
Dmitry Zarva
fa3566a9cb - fix: fmt issues 2026-04-17 16:20:16 +00:00
Dmitry Zarva
4e59e52454 - fix: ru docs 2026-04-17 14:10:20 +00:00
Dmitry Zarva
7b9b46291d - fix: param name in ru docs 2026-04-17 13:19:29 +00:00
Dmitry Zarva
2a168b2600 feat: make URLS to obtain proxy_secret, getProxyConfig, getProxyConfigV6 files optionally configurable 2026-04-17 13:04:46 +00:00
12 changed files with 633 additions and 185 deletions

View File

@@ -255,13 +255,22 @@ This document lists all configuration keys accepted by `config.toml`.
``` ```
## proxy_secret_path ## proxy_secret_path
- **Constraints / validation**: `String`. When omitted, the default path is `"proxy-secret"`. Empty values are accepted by TOML/serde but will likely fail at runtime (invalid file path). - **Constraints / validation**: `String`. When omitted, the default path is `"proxy-secret"`. Empty values are accepted by TOML/serde but will likely fail at runtime (invalid file path).
- **Description**: Path to Telegram infrastructure `proxy-secret` cache file used by ME handshake/RPC auth. Telemt always tries a fresh download from `https://core.telegram.org/getProxySecret` first, caches it to this path on success, and falls back to reading the cached file (any age) on download failure. - **Description**: Path to Telegram infrastructure `proxy-secret` cache file used by ME handshake/RPC auth. Telemt always tries a fresh download from `https://core.telegram.org/getProxySecret` first (unless `proxy_secret_url` is set) , caches it to this path on success, and falls back to reading the cached file (any age) on download failure.
- **Example**: - **Example**:
```toml ```toml
[general] [general]
proxy_secret_path = "proxy-secret" proxy_secret_path = "proxy-secret"
``` ```
## proxy_secret_url
- **Constraints / validation**: `String`. When omitted, the `"https://core.telegram.org/getProxySecret"` is used.
- **Description**: Optional URL to obtain `proxy-secret` file used by ME handshake/RPC auth. Telemt always tries a fresh download from this URL first (with fallback to `https://core.telegram.org/getProxySecret` if absent).
- **Example**:
```toml
[general]
proxy_secret_url = "https://core.telegram.org/getProxySecret"
```
## proxy_config_v4_cache_path ## proxy_config_v4_cache_path
- **Constraints / validation**: `String`. When set, must not be empty/whitespace-only. - **Constraints / validation**: `String`. When set, must not be empty/whitespace-only.
- **Description**: Optional disk cache path for raw `getProxyConfig` (IPv4) snapshot. At startup Telemt tries to fetch a fresh snapshot first; on fetch failure or empty snapshot it falls back to this cache file when present and non-empty. - **Description**: Optional disk cache path for raw `getProxyConfig` (IPv4) snapshot. At startup Telemt tries to fetch a fresh snapshot first; on fetch failure or empty snapshot it falls back to this cache file when present and non-empty.
@@ -271,6 +280,15 @@ This document lists all configuration keys accepted by `config.toml`.
[general] [general]
proxy_config_v4_cache_path = "cache/proxy-config-v4.txt" proxy_config_v4_cache_path = "cache/proxy-config-v4.txt"
``` ```
## proxy_config_v4_url
- **Constraints / validation**: `String`. When omitted, the `"https://core.telegram.org/getProxyConfig"` is used.
- **Description**: Optional URL to obtain raw `getProxyConfig` (IPv4). Telemt always tries a fresh download from this URL first (with fallback to `https://core.telegram.org/getProxyConfig` if absent).
- **Example**:
```toml
[general]
proxy_config_v4_url = "https://core.telegram.org/getProxyConfig"
```
## proxy_config_v6_cache_path ## proxy_config_v6_cache_path
- **Constraints / validation**: `String`. When set, must not be empty/whitespace-only. - **Constraints / validation**: `String`. When set, must not be empty/whitespace-only.
- **Description**: Optional disk cache path for raw `getProxyConfigV6` (IPv6) snapshot. At startup Telemt tries to fetch a fresh snapshot first; on fetch failure or empty snapshot it falls back to this cache file when present and non-empty. - **Description**: Optional disk cache path for raw `getProxyConfigV6` (IPv6) snapshot. At startup Telemt tries to fetch a fresh snapshot first; on fetch failure or empty snapshot it falls back to this cache file when present and non-empty.
@@ -280,6 +298,15 @@ This document lists all configuration keys accepted by `config.toml`.
[general] [general]
proxy_config_v6_cache_path = "cache/proxy-config-v6.txt" proxy_config_v6_cache_path = "cache/proxy-config-v6.txt"
``` ```
## proxy_config_v6_url
- **Constraints / validation**: `String`. When omitted, the `"https://core.telegram.org/getProxyConfigV6"` is used.
- **Description**: Optional URL to obtain raw `getProxyConfigV6` (IPv6). Telemt always tries a fresh download from this URL first (with fallback to `https://core.telegram.org/getProxyConfigV6` if absent).
- **Example**:
```toml
[general]
proxy_config_v6_url = "https://core.telegram.org/getProxyConfigV6"
```
## ad_tag ## ad_tag
- **Constraints / validation**: `String` (optional). When set, must be exactly 32 hex characters; invalid values are disabled during config load. - **Constraints / validation**: `String` (optional). When set, must be exactly 32 hex characters; invalid values are disabled during config load.
- **Description**: Global fallback sponsored-channel `ad_tag` (used when user has no override in `access.user_ad_tags`). An all-zero tag is accepted but has no effect (and is warned about) until replaced with a real tag from `@MTProxybot`. - **Description**: Global fallback sponsored-channel `ad_tag` (used when user has no override in `access.user_ad_tags`). An all-zero tag is accepted but has no effect (and is warned about) until replaced with a real tag from `@MTProxybot`.

View File

@@ -255,13 +255,22 @@
``` ```
## proxy_secret_path ## proxy_secret_path
- **Ограничения / валидация**: `String`. Если этот параметр не указан, используется путь по умолчанию — «proxy-secret». Пустые значения принимаются TOML/serde, но во время выполнения произойдет ошибка (invalid file path). - **Ограничения / валидация**: `String`. Если этот параметр не указан, используется путь по умолчанию — «proxy-secret». Пустые значения принимаются TOML/serde, но во время выполнения произойдет ошибка (invalid file path).
- **Описание**: Путь к файлу кэша `proxy-secret` инфраструктуры Telegram, используемому ME-handshake/аутентификацией RPC. Telemt всегда сначала пытается выполнить новую загрузку с https://core.telegram.org/getProxySecret, в случае успеха кэширует ее по этому пути и возвращается к чтению кэшированного файла в случае сбоя загрузки. - **Описание**: Путь к файлу кэша `proxy-secret` инфраструктуры Telegram, используемому ME-handshake/аутентификацией RPC. Telemt всегда сначала пытается выполнить новую загрузку с https://core.telegram.org/getProxySecret (если не установлен `proxy_secret_url`), в случае успеха кэширует ее по этому пути и возвращается к чтению кэшированного файла в случае сбоя загрузки.
- **Пример**: - **Пример**:
```toml ```toml
[general] [general]
proxy_secret_path = "proxy-secret" proxy_secret_path = "proxy-secret"
``` ```
## proxy_secret_url
- **Ограничения / валидация**: `String`. Если не указан, используется `"https://core.telegram.org/getProxySecret"`.
- **Описание**: Необязательный URL для получения файла `proxy-secret` используемого ME-handshake/аутентификацией RPC. Telemt всегда сначала пытается выполнить новую загрузку с этого URL (если не задан, используется https://core.telegram.org/getProxySecret).
- **Пример**:
```toml
[general]
proxy_secret_url = "https://core.telegram.org/getProxySecret"
```
## proxy_config_v4_cache_path ## proxy_config_v4_cache_path
- **Ограничения / валидация**: `String`. Если используется, значение не должно быть пустым или содержать только пробелы. - **Ограничения / валидация**: `String`. Если используется, значение не должно быть пустым или содержать только пробелы.
- **Описание**: Необязательный путь к кэшу для необработанного (raw) снимка getProxyConfig (IPv4). При запуске Telemt сначала пытается получить свежий снимок; в случае сбоя выборки или пустого снимка он возвращается к этому файлу кэша, если он присутствует и не пуст. - **Описание**: Необязательный путь к кэшу для необработанного (raw) снимка getProxyConfig (IPv4). При запуске Telemt сначала пытается получить свежий снимок; в случае сбоя выборки или пустого снимка он возвращается к этому файлу кэша, если он присутствует и не пуст.
@@ -271,6 +280,15 @@
[general] [general]
proxy_config_v4_cache_path = "cache/proxy-config-v4.txt" proxy_config_v4_cache_path = "cache/proxy-config-v4.txt"
``` ```
## proxy_config_v4_url
- **Ограничения / валидация**: `String`. Если не указан, используется `"https://core.telegram.org/getProxyConfig"`.
- **Описание**: Необязательный URL для получения `getProxyConfig` (IPv4). Telemt при всегда пытается выполнить новую загрузку с этого URL (и если не задан, использует `https://core.telegram.org/getProxyConfig`).
- **Example**:
```toml
[general]
proxy_config_v4_url = "https://core.telegram.org/getProxyConfig"
```
## proxy_config_v6_cache_path ## proxy_config_v6_cache_path
- **Ограничения / валидация**: `String`. Если используется, значение не должно быть пустым или содержать только пробелы. - **Ограничения / валидация**: `String`. Если используется, значение не должно быть пустым или содержать только пробелы.
- **Описание**: Необязательный путь к кэшу для необработанного (raw) снимка getProxyConfigV6 (IPv6). При запуске Telemt сначала пытается получить свежий снимок; в случае сбоя выборки или пустого снимка он возвращается к этому файлу кэша, если он присутствует и не пуст. - **Описание**: Необязательный путь к кэшу для необработанного (raw) снимка getProxyConfigV6 (IPv6). При запуске Telemt сначала пытается получить свежий снимок; в случае сбоя выборки или пустого снимка он возвращается к этому файлу кэша, если он присутствует и не пуст.
@@ -280,6 +298,15 @@
[general] [general]
proxy_config_v6_cache_path = "cache/proxy-config-v6.txt" proxy_config_v6_cache_path = "cache/proxy-config-v6.txt"
``` ```
## proxy_config_v6_url
- **Ограничения / валидация**: `String`. Если не указан, используется `"https://core.telegram.org/getProxyConfigV6"`.
- **Описание**: Необязательный URL для получения `getProxyConfigV6` (IPv6). Telemt при всегда пытается выполнить новую загрузку с этого URL (и если не задан, использует `https://core.telegram.org/getProxyConfigV6`).
- **Example**:
```toml
[general]
proxy_config_v6_url = "https://core.telegram.org/getProxyConfigV6"
```
## ad_tag ## ad_tag
- **Ограничения / валидация**: `String` (необязательный параметр). Если используется, значение должно быть ровно 32 символа в шестнадцатеричной системе; недопустимые значения отключаются во время загрузки конфигурации. - **Ограничения / валидация**: `String` (необязательный параметр). Если используется, значение должно быть ровно 32 символа в шестнадцатеричной системе; недопустимые значения отключаются во время загрузки конфигурации.
- **Описание**: Глобальный резервный спонсируемый канал `ad_tag` (используется, когда у пользователя нет переопределения в `access.user_ad_tags`). Тег со всеми нулями принимается, но не имеет никакого эффекта, пока не будет заменен реальным тегом от `@MTProxybot`. - **Описание**: Глобальный резервный спонсируемый канал `ad_tag` (используется, когда у пользователя нет переопределения в `access.user_ad_tags`). Тег со всеми нулями принимается, но не имеет никакого эффекта, пока не будет заменен реальным тегом от `@MTProxybot`.

View File

@@ -392,14 +392,26 @@ pub struct GeneralConfig {
#[serde(default = "default_proxy_secret_path")] #[serde(default = "default_proxy_secret_path")]
pub proxy_secret_path: Option<String>, pub proxy_secret_path: Option<String>,
/// Optional custom URL for infrastructure secret (https://core.telegram.org/getProxySecret if absent).
#[serde(default)]
pub proxy_secret_url: Option<String>,
/// Optional path to cache raw getProxyConfig (IPv4) snapshot for startup fallback. /// Optional path to cache raw getProxyConfig (IPv4) snapshot for startup fallback.
#[serde(default = "default_proxy_config_v4_cache_path")] #[serde(default = "default_proxy_config_v4_cache_path")]
pub proxy_config_v4_cache_path: Option<String>, pub proxy_config_v4_cache_path: Option<String>,
/// Optional custom URL for getProxyConfig (https://core.telegram.org/getProxyConfig if absent).
#[serde(default)]
pub proxy_config_v4_url: Option<String>,
/// Optional path to cache raw getProxyConfigV6 snapshot for startup fallback. /// Optional path to cache raw getProxyConfigV6 snapshot for startup fallback.
#[serde(default = "default_proxy_config_v6_cache_path")] #[serde(default = "default_proxy_config_v6_cache_path")]
pub proxy_config_v6_cache_path: Option<String>, pub proxy_config_v6_cache_path: Option<String>,
/// Optional custom URL for getProxyConfigV6 (https://core.telegram.org/getProxyConfigV6 if absent).
#[serde(default)]
pub proxy_config_v6_url: Option<String>,
/// Global ad_tag (32 hex chars from @MTProxybot). Fallback when user has no per-user tag in access.user_ad_tags. /// Global ad_tag (32 hex chars from @MTProxybot). Fallback when user has no per-user tag in access.user_ad_tags.
#[serde(default)] #[serde(default)]
pub ad_tag: Option<String>, pub ad_tag: Option<String>,
@@ -960,8 +972,11 @@ impl Default for GeneralConfig {
use_middle_proxy: default_true(), use_middle_proxy: default_true(),
ad_tag: None, ad_tag: None,
proxy_secret_path: default_proxy_secret_path(), proxy_secret_path: default_proxy_secret_path(),
proxy_secret_url: None,
proxy_config_v4_cache_path: default_proxy_config_v4_cache_path(), proxy_config_v4_cache_path: default_proxy_config_v4_cache_path(),
proxy_config_v4_url: None,
proxy_config_v6_cache_path: default_proxy_config_v6_cache_path(), proxy_config_v6_cache_path: default_proxy_config_v6_cache_path(),
proxy_config_v6_url: None,
middle_proxy_nat_ip: None, middle_proxy_nat_ip: None,
middle_proxy_nat_probe: default_true(), middle_proxy_nat_probe: default_true(),
middle_proxy_nat_stun: default_middle_proxy_nat_stun(), middle_proxy_nat_stun: default_middle_proxy_nat_stun(),

View File

@@ -66,6 +66,7 @@ pub(crate) async fn initialize_me_pool(
match crate::transport::middle_proxy::fetch_proxy_secret_with_upstream( match crate::transport::middle_proxy::fetch_proxy_secret_with_upstream(
proxy_secret_path, proxy_secret_path,
config.general.proxy_secret_len_max, config.general.proxy_secret_len_max,
config.general.proxy_secret_url.as_deref(),
Some(upstream_manager.clone()), Some(upstream_manager.clone()),
) )
.await .await
@@ -126,7 +127,11 @@ pub(crate) async fn initialize_me_pool(
.set_me_status(StartupMeStatus::Initializing, COMPONENT_ME_PROXY_CONFIG_V4) .set_me_status(StartupMeStatus::Initializing, COMPONENT_ME_PROXY_CONFIG_V4)
.await; .await;
let cfg_v4 = load_startup_proxy_config_snapshot( let cfg_v4 = load_startup_proxy_config_snapshot(
"https://core.telegram.org/getProxyConfig", config
.general
.proxy_config_v4_url
.as_deref()
.unwrap_or("https://core.telegram.org/getProxyConfig"),
config.general.proxy_config_v4_cache_path.as_deref(), config.general.proxy_config_v4_cache_path.as_deref(),
me2dc_fallback, me2dc_fallback,
"getProxyConfig", "getProxyConfig",
@@ -158,7 +163,11 @@ pub(crate) async fn initialize_me_pool(
.set_me_status(StartupMeStatus::Initializing, COMPONENT_ME_PROXY_CONFIG_V6) .set_me_status(StartupMeStatus::Initializing, COMPONENT_ME_PROXY_CONFIG_V6)
.await; .await;
let cfg_v6 = load_startup_proxy_config_snapshot( let cfg_v6 = load_startup_proxy_config_snapshot(
"https://core.telegram.org/getProxyConfigV6", config
.general
.proxy_config_v6_url
.as_deref()
.unwrap_or("https://core.telegram.org/getProxyConfigV6"),
config.general.proxy_config_v6_cache_path.as_deref(), config.general.proxy_config_v6_cache_path.as_deref(),
me2dc_fallback, me2dc_fallback,
"getProxyConfigV6", "getProxyConfigV6",

View File

@@ -321,7 +321,14 @@ async fn run_update_cycle(
let mut maps_changed = false; let mut maps_changed = false;
let mut ready_v4: Option<(ProxyConfigData, u64)> = None; let mut ready_v4: Option<(ProxyConfigData, u64)> = None;
let cfg_v4 = retry_fetch("https://core.telegram.org/getProxyConfig", upstream.clone()).await; let cfg_v4 = retry_fetch(
cfg.general
.proxy_config_v4_url
.as_deref()
.unwrap_or("https://core.telegram.org/getProxyConfig"),
upstream.clone(),
)
.await;
if let Some(cfg_v4) = cfg_v4 if let Some(cfg_v4) = cfg_v4
&& snapshot_passes_guards(cfg, &cfg_v4, "getProxyConfig") && snapshot_passes_guards(cfg, &cfg_v4, "getProxyConfig")
{ {
@@ -346,7 +353,10 @@ async fn run_update_cycle(
let mut ready_v6: Option<(ProxyConfigData, u64)> = None; let mut ready_v6: Option<(ProxyConfigData, u64)> = None;
let cfg_v6 = retry_fetch( let cfg_v6 = retry_fetch(
"https://core.telegram.org/getProxyConfigV6", cfg.general
.proxy_config_v6_url
.as_deref()
.unwrap_or("https://core.telegram.org/getProxyConfigV6"),
upstream.clone(), upstream.clone(),
) )
.await; .await;
@@ -430,6 +440,7 @@ async fn run_update_cycle(
match download_proxy_secret_with_max_len_via_upstream( match download_proxy_secret_with_max_len_via_upstream(
cfg.general.proxy_secret_len_max, cfg.general.proxy_secret_len_max,
upstream, upstream,
cfg.general.proxy_secret_url.as_deref(),
) )
.await .await
{ {

View File

@@ -7,7 +7,6 @@ mod model;
mod pressure; mod pressure;
mod scheduler; mod scheduler;
#[cfg(test)]
pub(crate) use model::PressureState; pub(crate) use model::PressureState;
pub(crate) use model::{AdmissionDecision, DispatchAction, DispatchFeedback, SchedulerDecision}; pub(crate) use model::{AdmissionDecision, DispatchAction, DispatchFeedback, SchedulerDecision};
pub(crate) use scheduler::{WorkerFairnessConfig, WorkerFairnessSnapshot, WorkerFairnessState}; pub(crate) use scheduler::{WorkerFairnessConfig, WorkerFairnessSnapshot, WorkerFairnessState};

View File

@@ -77,11 +77,12 @@ pub(crate) struct FlowFairnessState {
pub(crate) standing_state: StandingQueueState, pub(crate) standing_state: StandingQueueState,
pub(crate) scheduler_state: FlowSchedulerState, pub(crate) scheduler_state: FlowSchedulerState,
pub(crate) bucket_id: usize, pub(crate) bucket_id: usize,
pub(crate) weight_quanta: u8,
pub(crate) in_active_ring: bool, pub(crate) in_active_ring: bool,
} }
impl FlowFairnessState { impl FlowFairnessState {
pub(crate) fn new(flow_id: u64, worker_id: u16, bucket_id: usize) -> Self { pub(crate) fn new(flow_id: u64, worker_id: u16, bucket_id: usize, weight_quanta: u8) -> Self {
Self { Self {
_flow_id: flow_id, _flow_id: flow_id,
_worker_id: worker_id, _worker_id: worker_id,
@@ -97,6 +98,7 @@ impl FlowFairnessState {
standing_state: StandingQueueState::Transient, standing_state: StandingQueueState::Transient,
scheduler_state: FlowSchedulerState::Idle, scheduler_state: FlowSchedulerState::Idle,
bucket_id, bucket_id,
weight_quanta: weight_quanta.max(1),
in_active_ring: false, in_active_ring: false,
} }
} }

View File

@@ -146,59 +146,55 @@ impl PressureEvaluator {
((signals.standing_flows.saturating_mul(100)) / signals.active_flows).min(100) as u8 ((signals.standing_flows.saturating_mul(100)) / signals.active_flows).min(100) as u8
}; };
let mut pressure_score = 0u8; let mut pressured = false;
let mut saturated = false;
let queue_saturated_pct = cfg.queue_ratio_shedding_pct.min(cfg.queue_ratio_saturated_pct);
if queue_ratio_pct >= cfg.queue_ratio_pressured_pct { if queue_ratio_pct >= cfg.queue_ratio_pressured_pct {
pressure_score = pressure_score.max(1); pressured = true;
} }
if queue_ratio_pct >= cfg.queue_ratio_shedding_pct { if queue_ratio_pct >= queue_saturated_pct {
pressure_score = pressure_score.max(2); saturated = true;
}
if queue_ratio_pct >= cfg.queue_ratio_saturated_pct {
pressure_score = pressure_score.max(3);
} }
let standing_saturated_pct = cfg
.standing_ratio_shedding_pct
.min(cfg.standing_ratio_saturated_pct);
if standing_ratio_pct >= cfg.standing_ratio_pressured_pct { if standing_ratio_pct >= cfg.standing_ratio_pressured_pct {
pressure_score = pressure_score.max(1); pressured = true;
} }
if standing_ratio_pct >= cfg.standing_ratio_shedding_pct { if standing_ratio_pct >= standing_saturated_pct {
pressure_score = pressure_score.max(2); saturated = true;
}
if standing_ratio_pct >= cfg.standing_ratio_saturated_pct {
pressure_score = pressure_score.max(3);
} }
let rejects_saturated = cfg.rejects_shedding.min(cfg.rejects_saturated);
if self.admission_rejects_window >= cfg.rejects_pressured { if self.admission_rejects_window >= cfg.rejects_pressured {
pressure_score = pressure_score.max(1); pressured = true;
} }
if self.admission_rejects_window >= cfg.rejects_shedding { if self.admission_rejects_window >= rejects_saturated {
pressure_score = pressure_score.max(2); saturated = true;
}
if self.admission_rejects_window >= cfg.rejects_saturated {
pressure_score = pressure_score.max(3);
} }
let stalls_saturated = cfg.stalls_shedding.min(cfg.stalls_saturated);
if self.route_stalls_window >= cfg.stalls_pressured { if self.route_stalls_window >= cfg.stalls_pressured {
pressure_score = pressure_score.max(1); pressured = true;
} }
if self.route_stalls_window >= cfg.stalls_shedding { if self.route_stalls_window >= stalls_saturated {
pressure_score = pressure_score.max(2); saturated = true;
}
if self.route_stalls_window >= cfg.stalls_saturated {
pressure_score = pressure_score.max(3);
} }
if signals.backpressured_flows > signals.active_flows.saturating_div(2) if signals.backpressured_flows > signals.active_flows.saturating_div(2)
&& signals.active_flows > 0 && signals.active_flows > 0
{ {
pressure_score = pressure_score.max(2); pressured = true;
} }
match pressure_score { if saturated {
0 => PressureState::Normal, PressureState::Saturated
1 => PressureState::Pressured, } else if pressured {
2 => PressureState::Shedding, PressureState::Pressured
_ => PressureState::Saturated, } else {
PressureState::Normal
} }
} }

View File

@@ -1,7 +1,8 @@
use std::collections::{HashMap, VecDeque}; use std::collections::{HashMap, HashSet, VecDeque};
use std::time::{Duration, Instant}; use std::time::{Duration, Instant};
use bytes::Bytes; use bytes::Bytes;
use crate::protocol::constants::RPC_FLAG_QUICKACK;
use super::model::{ use super::model::{
AdmissionDecision, DispatchAction, DispatchCandidate, DispatchFeedback, FlowFairnessState, AdmissionDecision, DispatchAction, DispatchCandidate, DispatchFeedback, FlowFairnessState,
@@ -26,6 +27,8 @@ pub(crate) struct WorkerFairnessConfig {
pub(crate) max_consecutive_stalls_before_close: u8, pub(crate) max_consecutive_stalls_before_close: u8,
pub(crate) soft_bucket_count: usize, pub(crate) soft_bucket_count: usize,
pub(crate) soft_bucket_share_pct: u8, pub(crate) soft_bucket_share_pct: u8,
pub(crate) default_flow_weight: u8,
pub(crate) quickack_flow_weight: u8,
pub(crate) pressure: PressureConfig, pub(crate) pressure: PressureConfig,
} }
@@ -46,6 +49,8 @@ impl Default for WorkerFairnessConfig {
max_consecutive_stalls_before_close: 16, max_consecutive_stalls_before_close: 16,
soft_bucket_count: 64, soft_bucket_count: 64,
soft_bucket_share_pct: 25, soft_bucket_share_pct: 25,
default_flow_weight: 1,
quickack_flow_weight: 4,
pressure: PressureConfig::default(), pressure: PressureConfig::default(),
} }
} }
@@ -57,9 +62,9 @@ struct FlowEntry {
} }
impl FlowEntry { impl FlowEntry {
fn new(flow_id: u64, worker_id: u16, bucket_id: usize) -> Self { fn new(flow_id: u64, worker_id: u16, bucket_id: usize, weight_quanta: u8) -> Self {
Self { Self {
fairness: FlowFairnessState::new(flow_id, worker_id, bucket_id), fairness: FlowFairnessState::new(flow_id, worker_id, bucket_id, weight_quanta),
queue: VecDeque::new(), queue: VecDeque::new(),
} }
} }
@@ -86,6 +91,7 @@ pub(crate) struct WorkerFairnessState {
pressure: PressureEvaluator, pressure: PressureEvaluator,
flows: HashMap<u64, FlowEntry>, flows: HashMap<u64, FlowEntry>,
active_ring: VecDeque<u64>, active_ring: VecDeque<u64>,
active_ring_members: HashSet<u64>,
total_queued_bytes: u64, total_queued_bytes: u64,
bucket_queued_bytes: Vec<u64>, bucket_queued_bytes: Vec<u64>,
bucket_active_flows: Vec<usize>, bucket_active_flows: Vec<usize>,
@@ -108,6 +114,7 @@ impl WorkerFairnessState {
pressure: PressureEvaluator::new(now), pressure: PressureEvaluator::new(now),
flows: HashMap::new(), flows: HashMap::new(),
active_ring: VecDeque::new(), active_ring: VecDeque::new(),
active_ring_members: HashSet::new(),
total_queued_bytes: 0, total_queued_bytes: 0,
bucket_queued_bytes: vec![0; bucket_count], bucket_queued_bytes: vec![0; bucket_count],
bucket_active_flows: vec![0; bucket_count], bucket_active_flows: vec![0; bucket_count],
@@ -184,6 +191,7 @@ impl WorkerFairnessState {
} }
let bucket_id = self.bucket_for(conn_id); let bucket_id = self.bucket_for(conn_id);
let frame_weight = Self::weight_for_flags(&self.config, flags);
let bucket_cap = self let bucket_cap = self
.config .config
.max_total_queued_bytes .max_total_queued_bytes
@@ -205,12 +213,13 @@ impl WorkerFairnessState {
self.bucket_active_flows[bucket_id].saturating_add(1); self.bucket_active_flows[bucket_id].saturating_add(1);
self.flows.insert( self.flows.insert(
conn_id, conn_id,
FlowEntry::new(conn_id, self.config.worker_id, bucket_id), FlowEntry::new(conn_id, self.config.worker_id, bucket_id, frame_weight),
); );
self.flows self.flows
.get_mut(&conn_id) .get_mut(&conn_id)
.expect("flow inserted must be retrievable") .expect("flow inserted must be retrievable")
}; };
entry.fairness.weight_quanta = entry.fairness.weight_quanta.max(frame_weight);
if entry.fairness.pending_bytes.saturating_add(frame_bytes) if entry.fairness.pending_bytes.saturating_add(frame_bytes)
> self.config.max_flow_queued_bytes > self.config.max_flow_queued_bytes
@@ -242,11 +251,24 @@ impl WorkerFairnessState {
self.bucket_queued_bytes[bucket_id] = self.bucket_queued_bytes[bucket_id] =
self.bucket_queued_bytes[bucket_id].saturating_add(frame_bytes); self.bucket_queued_bytes[bucket_id].saturating_add(frame_bytes);
let mut enqueue_active = false;
if !entry.fairness.in_active_ring { if !entry.fairness.in_active_ring {
entry.fairness.in_active_ring = true; entry.fairness.in_active_ring = true;
self.active_ring.push_back(conn_id); enqueue_active = true;
} }
let pressure_state = self.pressure.state();
let (before_membership, after_membership) = {
let before = Self::flow_membership(&entry.fairness);
Self::classify_flow(&self.config, pressure_state, now, &mut entry.fairness);
let after = Self::flow_membership(&entry.fairness);
(before, after)
};
if enqueue_active {
self.enqueue_active_conn(conn_id);
}
self.apply_flow_membership_delta(before_membership, after_membership);
self.evaluate_pressure(now, true); self.evaluate_pressure(now, true);
AdmissionDecision::Admit AdmissionDecision::Admit
} }
@@ -260,32 +282,44 @@ impl WorkerFairnessState {
let Some(conn_id) = self.active_ring.pop_front() else { let Some(conn_id) = self.active_ring.pop_front() else {
break; break;
}; };
if !self.active_ring_members.remove(&conn_id) {
continue;
}
let mut candidate = None; let mut candidate = None;
let mut requeue_active = false; let mut requeue_active = false;
let mut drained_bytes = 0u64; let mut drained_bytes = 0u64;
let mut bucket_id = 0usize; let mut bucket_id = 0usize;
let mut should_continue = false;
let mut enqueue_active = false;
let mut membership_delta = None;
let pressure_state = self.pressure.state(); let pressure_state = self.pressure.state();
if let Some(flow) = self.flows.get_mut(&conn_id) { if let Some(flow) = self.flows.get_mut(&conn_id) {
bucket_id = flow.fairness.bucket_id; bucket_id = flow.fairness.bucket_id;
flow.fairness.in_active_ring = false;
let before_membership = Self::flow_membership(&flow.fairness);
if flow.queue.is_empty() { if flow.queue.is_empty() {
flow.fairness.in_active_ring = false; flow.fairness.in_active_ring = false;
flow.fairness.scheduler_state = FlowSchedulerState::Idle; flow.fairness.scheduler_state = FlowSchedulerState::Idle;
flow.fairness.pending_bytes = 0; flow.fairness.pending_bytes = 0;
flow.fairness.deficit_bytes = 0;
flow.fairness.queue_started_at = None; flow.fairness.queue_started_at = None;
continue; should_continue = true;
} } else {
Self::classify_flow(&self.config, pressure_state, now, &mut flow.fairness); Self::classify_flow(&self.config, pressure_state, now, &mut flow.fairness);
let quantum = let quantum = Self::effective_quantum_bytes(
Self::effective_quantum_bytes(&self.config, pressure_state, &flow.fairness); &self.config,
pressure_state,
&flow.fairness,
);
flow.fairness.deficit_bytes = flow flow.fairness.deficit_bytes = flow
.fairness .fairness
.deficit_bytes .deficit_bytes
.saturating_add(i64::from(quantum)); .saturating_add(i64::from(quantum));
Self::clamp_deficit_bytes(&self.config, &mut flow.fairness);
self.deficit_grants = self.deficit_grants.saturating_add(1); self.deficit_grants = self.deficit_grants.saturating_add(1);
let front_len = flow.queue.front().map_or(0, |front| front.queued_bytes()); let front_len = flow.queue.front().map_or(0, |front| front.queued_bytes());
@@ -294,6 +328,8 @@ impl WorkerFairnessState {
flow.fairness.consecutive_skips.saturating_add(1); flow.fairness.consecutive_skips.saturating_add(1);
self.deficit_skips = self.deficit_skips.saturating_add(1); self.deficit_skips = self.deficit_skips.saturating_add(1);
requeue_active = true; requeue_active = true;
flow.fairness.in_active_ring = true;
enqueue_active = true;
} else if let Some(frame) = flow.queue.pop_front() { } else if let Some(frame) = flow.queue.pop_front() {
drained_bytes = frame.queued_bytes(); drained_bytes = frame.queued_bytes();
flow.fairness.pending_bytes = flow.fairness.pending_bytes =
@@ -302,6 +338,7 @@ impl WorkerFairnessState {
.fairness .fairness
.deficit_bytes .deficit_bytes
.saturating_sub(drained_bytes as i64); .saturating_sub(drained_bytes as i64);
Self::clamp_deficit_bytes(&self.config, &mut flow.fairness);
flow.fairness.consecutive_skips = 0; flow.fairness.consecutive_skips = 0;
flow.fairness.queue_started_at = flow.fairness.queue_started_at =
flow.queue.front().map(|front| front.enqueued_at); flow.queue.front().map(|front| front.enqueued_at);
@@ -309,6 +346,10 @@ impl WorkerFairnessState {
if !requeue_active { if !requeue_active {
flow.fairness.scheduler_state = FlowSchedulerState::Idle; flow.fairness.scheduler_state = FlowSchedulerState::Idle;
flow.fairness.in_active_ring = false; flow.fairness.in_active_ring = false;
flow.fairness.deficit_bytes = 0;
} else {
flow.fairness.in_active_ring = true;
enqueue_active = true;
} }
candidate = Some(DispatchCandidate { candidate = Some(DispatchCandidate {
pressure_state, pressure_state,
@@ -318,17 +359,25 @@ impl WorkerFairnessState {
} }
} }
membership_delta = Some((before_membership, Self::flow_membership(&flow.fairness)));
}
if let Some((before_membership, after_membership)) = membership_delta {
self.apply_flow_membership_delta(before_membership, after_membership);
}
if should_continue {
continue;
}
if drained_bytes > 0 { if drained_bytes > 0 {
self.total_queued_bytes = self.total_queued_bytes.saturating_sub(drained_bytes); self.total_queued_bytes = self.total_queued_bytes.saturating_sub(drained_bytes);
self.bucket_queued_bytes[bucket_id] = self.bucket_queued_bytes[bucket_id] =
self.bucket_queued_bytes[bucket_id].saturating_sub(drained_bytes); self.bucket_queued_bytes[bucket_id].saturating_sub(drained_bytes);
} }
if requeue_active { if requeue_active && enqueue_active {
if let Some(flow) = self.flows.get_mut(&conn_id) { self.enqueue_active_conn(conn_id);
flow.fairness.in_active_ring = true;
}
self.active_ring.push_back(conn_id);
} }
if let Some(candidate) = candidate { if let Some(candidate) = candidate {
@@ -348,7 +397,9 @@ impl WorkerFairnessState {
) -> DispatchAction { ) -> DispatchAction {
match feedback { match feedback {
DispatchFeedback::Routed => { DispatchFeedback::Routed => {
let mut membership_delta = None;
if let Some(flow) = self.flows.get_mut(&conn_id) { if let Some(flow) = self.flows.get_mut(&conn_id) {
let before_membership = Self::flow_membership(&flow.fairness);
flow.fairness.last_drain_at = Some(now); flow.fairness.last_drain_at = Some(now);
flow.fairness.recent_drain_bytes = flow flow.fairness.recent_drain_bytes = flow
.fairness .fairness
@@ -358,6 +409,17 @@ impl WorkerFairnessState {
if flow.fairness.scheduler_state != FlowSchedulerState::Idle { if flow.fairness.scheduler_state != FlowSchedulerState::Idle {
flow.fairness.scheduler_state = FlowSchedulerState::Active; flow.fairness.scheduler_state = FlowSchedulerState::Active;
} }
Self::classify_flow(
&self.config,
self.pressure.state(),
now,
&mut flow.fairness,
);
membership_delta =
Some((before_membership, Self::flow_membership(&flow.fairness)));
}
if let Some((before_membership, after_membership)) = membership_delta {
self.apply_flow_membership_delta(before_membership, after_membership);
} }
self.evaluate_pressure(now, false); self.evaluate_pressure(now, false);
DispatchAction::Continue DispatchAction::Continue
@@ -365,17 +427,20 @@ impl WorkerFairnessState {
DispatchFeedback::QueueFull => { DispatchFeedback::QueueFull => {
self.pressure.note_route_stall(now, &self.config.pressure); self.pressure.note_route_stall(now, &self.config.pressure);
self.downstream_stalls = self.downstream_stalls.saturating_add(1); self.downstream_stalls = self.downstream_stalls.saturating_add(1);
let state = self.pressure.state();
let Some(flow) = self.flows.get_mut(&conn_id) else { let Some(flow) = self.flows.get_mut(&conn_id) else {
self.evaluate_pressure(now, true); self.evaluate_pressure(now, true);
return DispatchAction::Continue; return DispatchAction::Continue;
}; };
let (before_membership, after_membership, should_close_flow, enqueue_active) = {
let before_membership = Self::flow_membership(&flow.fairness);
let mut enqueue_active = false;
flow.fairness.consecutive_stalls = flow.fairness.consecutive_stalls =
flow.fairness.consecutive_stalls.saturating_add(1); flow.fairness.consecutive_stalls.saturating_add(1);
flow.fairness.scheduler_state = FlowSchedulerState::Backpressured; flow.fairness.scheduler_state = FlowSchedulerState::Backpressured;
flow.fairness.pressure_class = FlowPressureClass::Backpressured; flow.fairness.pressure_class = FlowPressureClass::Backpressured;
let state = self.pressure.state();
let should_shed_frame = matches!(state, PressureState::Saturated) let should_shed_frame = matches!(state, PressureState::Saturated)
|| (matches!(state, PressureState::Shedding) || (matches!(state, PressureState::Shedding)
&& flow.fairness.standing_state == StandingQueueState::Standing && flow.fairness.standing_state == StandingQueueState::Standing
@@ -392,20 +457,35 @@ impl WorkerFairnessState {
flow.fairness.pending_bytes.saturating_add(frame_bytes); flow.fairness.pending_bytes.saturating_add(frame_bytes);
flow.fairness.queue_started_at = flow.fairness.queue_started_at =
flow.queue.front().map(|front| front.enqueued_at); flow.queue.front().map(|front| front.enqueued_at);
self.total_queued_bytes = self.total_queued_bytes.saturating_add(frame_bytes); self.total_queued_bytes =
self.bucket_queued_bytes[flow.fairness.bucket_id] = self.bucket_queued_bytes self.total_queued_bytes.saturating_add(frame_bytes);
[flow.fairness.bucket_id] self.bucket_queued_bytes[flow.fairness.bucket_id] = self
.bucket_queued_bytes[flow.fairness.bucket_id]
.saturating_add(frame_bytes); .saturating_add(frame_bytes);
if !flow.fairness.in_active_ring { if !flow.fairness.in_active_ring {
flow.fairness.in_active_ring = true; flow.fairness.in_active_ring = true;
self.active_ring.push_back(conn_id); enqueue_active = true;
} }
} }
if flow.fairness.consecutive_stalls Self::classify_flow(&self.config, state, now, &mut flow.fairness);
let after_membership = Self::flow_membership(&flow.fairness);
let should_close_flow = flow.fairness.consecutive_stalls
>= self.config.max_consecutive_stalls_before_close >= self.config.max_consecutive_stalls_before_close
&& self.pressure.state() == PressureState::Saturated && self.pressure.state() == PressureState::Saturated;
{ (
before_membership,
after_membership,
should_close_flow,
enqueue_active,
)
};
if enqueue_active {
self.enqueue_active_conn(conn_id);
}
self.apply_flow_membership_delta(before_membership, after_membership);
if should_close_flow {
self.remove_flow(conn_id); self.remove_flow(conn_id);
self.evaluate_pressure(now, true); self.evaluate_pressure(now, true);
return DispatchAction::CloseFlow; return DispatchAction::CloseFlow;
@@ -426,6 +506,15 @@ impl WorkerFairnessState {
let Some(entry) = self.flows.remove(&conn_id) else { let Some(entry) = self.flows.remove(&conn_id) else {
return; return;
}; };
self.active_ring_members.remove(&conn_id);
self.active_ring.retain(|queued_conn_id| *queued_conn_id != conn_id);
let (was_standing, was_backpressured) = Self::flow_membership(&entry.fairness);
if was_standing {
self.standing_flow_count = self.standing_flow_count.saturating_sub(1);
}
if was_backpressured {
self.backpressured_flow_count = self.backpressured_flow_count.saturating_sub(1);
}
self.bucket_active_flows[entry.fairness.bucket_id] = self.bucket_active_flows[entry.fairness.bucket_id] =
self.bucket_active_flows[entry.fairness.bucket_id].saturating_sub(1); self.bucket_active_flows[entry.fairness.bucket_id].saturating_sub(1);
@@ -440,27 +529,6 @@ impl WorkerFairnessState {
} }
fn evaluate_pressure(&mut self, now: Instant, force: bool) { fn evaluate_pressure(&mut self, now: Instant, force: bool) {
let mut standing = 0usize;
let mut backpressured = 0usize;
for flow in self.flows.values_mut() {
Self::classify_flow(&self.config, self.pressure.state(), now, &mut flow.fairness);
if flow.fairness.standing_state == StandingQueueState::Standing {
standing = standing.saturating_add(1);
}
if matches!(
flow.fairness.scheduler_state,
FlowSchedulerState::Backpressured
| FlowSchedulerState::Penalized
| FlowSchedulerState::SheddingCandidate
) {
backpressured = backpressured.saturating_add(1);
}
}
self.standing_flow_count = standing;
self.backpressured_flow_count = backpressured;
let _ = self.pressure.maybe_evaluate( let _ = self.pressure.maybe_evaluate(
now, now,
&self.config.pressure, &self.config.pressure,
@@ -468,8 +536,8 @@ impl WorkerFairnessState {
PressureSignals { PressureSignals {
active_flows: self.flows.len(), active_flows: self.flows.len(),
total_queued_bytes: self.total_queued_bytes, total_queued_bytes: self.total_queued_bytes,
standing_flows: standing, standing_flows: self.standing_flow_count,
backpressured_flows: backpressured, backpressured_flows: self.backpressured_flow_count,
}, },
force, force,
); );
@@ -481,12 +549,39 @@ impl WorkerFairnessState {
now: Instant, now: Instant,
fairness: &mut FlowFairnessState, fairness: &mut FlowFairnessState,
) { ) {
if fairness.pending_bytes == 0 { let (pressure_class, standing_state, scheduler_state, standing) =
fairness.pressure_class = FlowPressureClass::Healthy; Self::derive_flow_classification(config, pressure_state, now, fairness);
fairness.standing_state = StandingQueueState::Transient; fairness.pressure_class = pressure_class;
fairness.scheduler_state = FlowSchedulerState::Idle; fairness.standing_state = standing_state;
fairness.scheduler_state = scheduler_state;
if scheduler_state == FlowSchedulerState::Idle {
fairness.deficit_bytes = 0;
}
if standing {
fairness.penalty_score = fairness.penalty_score.saturating_add(1);
} else {
fairness.penalty_score = fairness.penalty_score.saturating_sub(1); fairness.penalty_score = fairness.penalty_score.saturating_sub(1);
return; }
}
fn derive_flow_classification(
config: &WorkerFairnessConfig,
pressure_state: PressureState,
now: Instant,
fairness: &FlowFairnessState,
) -> (
FlowPressureClass,
StandingQueueState,
FlowSchedulerState,
bool,
) {
if fairness.pending_bytes == 0 {
return (
FlowPressureClass::Healthy,
StandingQueueState::Transient,
FlowSchedulerState::Idle,
false,
);
} }
let queue_age = fairness let queue_age = fairness
@@ -503,29 +598,165 @@ impl WorkerFairnessState {
&& (fairness.consecutive_stalls >= config.standing_stall_threshold || drain_stalled); && (fairness.consecutive_stalls >= config.standing_stall_threshold || drain_stalled);
if standing { if standing {
fairness.standing_state = StandingQueueState::Standing; let scheduler_state = if pressure_state >= PressureState::Shedding {
fairness.pressure_class = FlowPressureClass::Standing;
fairness.penalty_score = fairness.penalty_score.saturating_add(1);
fairness.scheduler_state = if pressure_state >= PressureState::Shedding {
FlowSchedulerState::SheddingCandidate FlowSchedulerState::SheddingCandidate
} else { } else {
FlowSchedulerState::Penalized FlowSchedulerState::Penalized
}; };
return; return (
FlowPressureClass::Standing,
StandingQueueState::Standing,
scheduler_state,
true,
);
} }
fairness.standing_state = StandingQueueState::Transient;
if fairness.consecutive_stalls > 0 { if fairness.consecutive_stalls > 0 {
fairness.pressure_class = FlowPressureClass::Backpressured; return (
fairness.scheduler_state = FlowSchedulerState::Backpressured; FlowPressureClass::Backpressured,
} else if fairness.pending_bytes >= config.standing_queue_min_backlog_bytes { StandingQueueState::Transient,
fairness.pressure_class = FlowPressureClass::Bursty; FlowSchedulerState::Backpressured,
fairness.scheduler_state = FlowSchedulerState::Active; false,
} else { );
fairness.pressure_class = FlowPressureClass::Healthy;
fairness.scheduler_state = FlowSchedulerState::Active;
} }
fairness.penalty_score = fairness.penalty_score.saturating_sub(1);
if fairness.pending_bytes >= config.standing_queue_min_backlog_bytes {
return (
FlowPressureClass::Bursty,
StandingQueueState::Transient,
FlowSchedulerState::Active,
false,
);
}
(
FlowPressureClass::Healthy,
StandingQueueState::Transient,
FlowSchedulerState::Active,
false,
)
}
#[inline]
fn flow_membership(fairness: &FlowFairnessState) -> (bool, bool) {
(
fairness.standing_state == StandingQueueState::Standing,
Self::scheduler_state_is_backpressured(fairness.scheduler_state),
)
}
#[inline]
fn scheduler_state_is_backpressured(state: FlowSchedulerState) -> bool {
matches!(
state,
FlowSchedulerState::Backpressured
| FlowSchedulerState::Penalized
| FlowSchedulerState::SheddingCandidate
)
}
fn apply_flow_membership_delta(
&mut self,
before_membership: (bool, bool),
after_membership: (bool, bool),
) {
if before_membership.0 != after_membership.0 {
if after_membership.0 {
self.standing_flow_count = self.standing_flow_count.saturating_add(1);
} else {
self.standing_flow_count = self.standing_flow_count.saturating_sub(1);
}
}
if before_membership.1 != after_membership.1 {
if after_membership.1 {
self.backpressured_flow_count = self.backpressured_flow_count.saturating_add(1);
} else {
self.backpressured_flow_count = self.backpressured_flow_count.saturating_sub(1);
}
}
}
#[inline]
fn clamp_deficit_bytes(config: &WorkerFairnessConfig, fairness: &mut FlowFairnessState) {
let max_deficit = config.max_flow_queued_bytes.min(i64::MAX as u64) as i64;
fairness.deficit_bytes = fairness.deficit_bytes.clamp(0, max_deficit);
}
#[inline]
fn enqueue_active_conn(&mut self, conn_id: u64) {
if self.active_ring_members.insert(conn_id) {
self.active_ring.push_back(conn_id);
}
}
#[inline]
fn weight_for_flags(config: &WorkerFairnessConfig, flags: u32) -> u8 {
if (flags & RPC_FLAG_QUICKACK) != 0 {
return config.quickack_flow_weight.max(1);
}
config.default_flow_weight.max(1)
}
#[cfg(test)]
pub(crate) fn debug_recompute_flow_counters(&self, now: Instant) -> (usize, usize) {
let pressure_state = self.pressure.state();
let mut standing = 0usize;
let mut backpressured = 0usize;
for flow in self.flows.values() {
let (_, standing_state, scheduler_state, _) =
Self::derive_flow_classification(&self.config, pressure_state, now, &flow.fairness);
if standing_state == StandingQueueState::Standing {
standing = standing.saturating_add(1);
}
if Self::scheduler_state_is_backpressured(scheduler_state) {
backpressured = backpressured.saturating_add(1);
}
}
(standing, backpressured)
}
#[cfg(test)]
pub(crate) fn debug_check_active_ring_consistency(&self) -> bool {
if self.active_ring.len() != self.active_ring_members.len() {
return false;
}
let mut seen = HashSet::with_capacity(self.active_ring.len());
for conn_id in self.active_ring.iter().copied() {
if !seen.insert(conn_id) {
return false;
}
if !self.active_ring_members.contains(&conn_id) {
return false;
}
let Some(flow) = self.flows.get(&conn_id) else {
return false;
};
if !flow.fairness.in_active_ring || flow.queue.is_empty() {
return false;
}
}
for (conn_id, flow) in self.flows.iter() {
let in_ring = self.active_ring_members.contains(conn_id);
if flow.fairness.in_active_ring != in_ring {
return false;
}
if in_ring && flow.queue.is_empty() {
return false;
}
}
true
}
#[cfg(test)]
pub(crate) fn debug_max_deficit_bytes(&self) -> i64 {
self.flows
.values()
.map(|entry| entry.fairness.deficit_bytes)
.max()
.unwrap_or(0)
} }
fn effective_quantum_bytes( fn effective_quantum_bytes(
@@ -542,12 +773,14 @@ impl WorkerFairnessState {
return config.penalized_quantum_bytes.max(1); return config.penalized_quantum_bytes.max(1);
} }
match pressure_state { let base_quantum = match pressure_state {
PressureState::Normal => config.base_quantum_bytes.max(1), PressureState::Normal => config.base_quantum_bytes.max(1),
PressureState::Pressured => config.pressured_quantum_bytes.max(1), PressureState::Pressured => config.pressured_quantum_bytes.max(1),
PressureState::Shedding => config.pressured_quantum_bytes.max(1), PressureState::Shedding => config.pressured_quantum_bytes.max(1),
PressureState::Saturated => config.penalized_quantum_bytes.max(1), PressureState::Saturated => config.penalized_quantum_bytes.max(1),
} };
let weighted_quantum = base_quantum.saturating_mul(fairness.weight_quanta.max(1) as u32);
weighted_quantum.max(1)
} }
fn bucket_for(&self, conn_id: u64) -> usize { fn bucket_for(&self, conn_id: u64) -> usize {

View File

@@ -4,7 +4,7 @@ use std::collections::HashMap;
use std::io::ErrorKind; use std::io::ErrorKind;
use std::sync::Arc; use std::sync::Arc;
use std::sync::atomic::{AtomicBool, AtomicU32, AtomicU64, Ordering}; use std::sync::atomic::{AtomicBool, AtomicU32, AtomicU64, Ordering};
use std::time::Instant; use std::time::{Duration, Instant};
use bytes::{Bytes, BytesMut}; use bytes::{Bytes, BytesMut};
use tokio::io::AsyncReadExt; use tokio::io::AsyncReadExt;
@@ -21,8 +21,8 @@ use crate::stats::Stats;
use super::codec::{RpcChecksumMode, WriterCommand, rpc_crc}; use super::codec::{RpcChecksumMode, WriterCommand, rpc_crc};
use super::fairness::{ use super::fairness::{
AdmissionDecision, DispatchAction, DispatchFeedback, SchedulerDecision, WorkerFairnessConfig, AdmissionDecision, DispatchAction, DispatchFeedback, PressureState, SchedulerDecision,
WorkerFairnessSnapshot, WorkerFairnessState, WorkerFairnessConfig, WorkerFairnessSnapshot, WorkerFairnessState,
}; };
use super::registry::RouteResult; use super::registry::RouteResult;
use super::{ConnRegistry, MeResponse}; use super::{ConnRegistry, MeResponse};
@@ -45,10 +45,22 @@ fn is_data_route_queue_full(result: RouteResult) -> bool {
) )
} }
fn should_close_on_queue_full_streak(streak: u8) -> bool { fn should_close_on_queue_full_streak(streak: u8, pressure_state: PressureState) -> bool {
if pressure_state < PressureState::Shedding {
return false;
}
streak >= DATA_ROUTE_QUEUE_FULL_STARVATION_THRESHOLD streak >= DATA_ROUTE_QUEUE_FULL_STARVATION_THRESHOLD
} }
fn should_schedule_fairness_retry(snapshot: &WorkerFairnessSnapshot) -> bool {
snapshot.total_queued_bytes > 0
}
fn fairness_retry_delay(route_wait_ms: u64) -> Duration {
Duration::from_millis(route_wait_ms.max(1))
}
async fn route_data_with_retry( async fn route_data_with_retry(
reg: &ConnRegistry, reg: &ConnRegistry,
conn_id: u64, conn_id: u64,
@@ -157,7 +169,7 @@ async fn drain_fairness_scheduler(
break; break;
}; };
let cid = candidate.frame.conn_id; let cid = candidate.frame.conn_id;
let _pressure_state = candidate.pressure_state; let pressure_state = candidate.pressure_state;
let _flow_class = candidate.flow_class; let _flow_class = candidate.flow_class;
let routed = route_data_with_retry( let routed = route_data_with_retry(
reg, reg,
@@ -176,7 +188,7 @@ async fn drain_fairness_scheduler(
if is_data_route_queue_full(routed) { if is_data_route_queue_full(routed) {
let streak = data_route_queue_full_streak.entry(cid).or_insert(0); let streak = data_route_queue_full_streak.entry(cid).or_insert(0);
*streak = streak.saturating_add(1); *streak = streak.saturating_add(1);
if should_close_on_queue_full_streak(*streak) { if should_close_on_queue_full_streak(*streak, pressure_state) {
fairness.remove_flow(cid); fairness.remove_flow(cid);
data_route_queue_full_streak.remove(&cid); data_route_queue_full_streak.remove(&cid);
reg.unregister(cid).await; reg.unregister(cid).await;
@@ -231,10 +243,33 @@ pub(crate) async fn reader_loop(
let mut fairness_snapshot = fairness.snapshot(); let mut fairness_snapshot = fairness.snapshot();
loop { loop {
let mut tmp = [0u8; 65_536]; let mut tmp = [0u8; 65_536];
let backlog_retry_enabled = should_schedule_fairness_retry(&fairness_snapshot);
let backlog_retry_delay =
fairness_retry_delay(reader_route_data_wait_ms.load(Ordering::Relaxed));
let mut retry_only = false;
let n = tokio::select! { let n = tokio::select! {
res = rd.read(&mut tmp) => res.map_err(ProxyError::Io)?, res = rd.read(&mut tmp) => res.map_err(ProxyError::Io)?,
_ = tokio::time::sleep(backlog_retry_delay), if backlog_retry_enabled => {
retry_only = true;
0usize
},
_ = cancel.cancelled() => return Ok(()), _ = cancel.cancelled() => return Ok(()),
}; };
if retry_only {
let route_wait_ms = reader_route_data_wait_ms.load(Ordering::Relaxed);
drain_fairness_scheduler(
&mut fairness,
reg.as_ref(),
&tx,
&mut data_route_queue_full_streak,
route_wait_ms,
stats.as_ref(),
)
.await;
let current_snapshot = fairness.snapshot();
apply_fairness_metrics_delta(stats.as_ref(), &mut fairness_snapshot, current_snapshot);
continue;
}
if n == 0 { if n == 0 {
stats.increment_me_reader_eof_total(); stats.increment_me_reader_eof_total();
return Err(ProxyError::Io(std::io::Error::new( return Err(ProxyError::Io(std::io::Error::new(
@@ -317,12 +352,9 @@ pub(crate) async fn reader_loop(
stats.increment_me_route_drop_queue_full_high(); stats.increment_me_route_drop_queue_full_high();
let streak = data_route_queue_full_streak.entry(cid).or_insert(0); let streak = data_route_queue_full_streak.entry(cid).or_insert(0);
*streak = streak.saturating_add(1); *streak = streak.saturating_add(1);
if should_close_on_queue_full_streak(*streak) let pressure_state = fairness.pressure_state();
|| matches!( if should_close_on_queue_full_streak(*streak, pressure_state)
admission, || matches!(admission, AdmissionDecision::RejectSaturated)
AdmissionDecision::RejectSaturated
| AdmissionDecision::RejectStandingFlow
)
{ {
fairness.remove_flow(cid); fairness.remove_flow(cid);
data_route_queue_full_streak.remove(&cid); data_route_queue_full_streak.remove(&cid);
@@ -445,14 +477,18 @@ pub(crate) async fn reader_loop(
#[cfg(test)] #[cfg(test)]
mod tests { mod tests {
use std::time::Duration;
use bytes::Bytes; use bytes::Bytes;
use super::PressureState;
use crate::transport::middle_proxy::ConnRegistry; use crate::transport::middle_proxy::ConnRegistry;
use super::{ use super::{
MeResponse, RouteResult, is_data_route_queue_full, route_data_with_retry, MeResponse, RouteResult, WorkerFairnessSnapshot, fairness_retry_delay,
should_close_on_queue_full_streak, should_close_on_route_result_for_ack, is_data_route_queue_full, route_data_with_retry, should_close_on_queue_full_streak,
should_close_on_route_result_for_data, should_close_on_route_result_for_ack, should_close_on_route_result_for_data,
should_schedule_fairness_retry,
}; };
#[test] #[test]
@@ -475,10 +511,29 @@ mod tests {
assert!(is_data_route_queue_full(RouteResult::QueueFullBase)); assert!(is_data_route_queue_full(RouteResult::QueueFullBase));
assert!(is_data_route_queue_full(RouteResult::QueueFullHigh)); assert!(is_data_route_queue_full(RouteResult::QueueFullHigh));
assert!(!is_data_route_queue_full(RouteResult::NoConn)); assert!(!is_data_route_queue_full(RouteResult::NoConn));
assert!(!should_close_on_queue_full_streak(1)); assert!(!should_close_on_queue_full_streak(1, PressureState::Normal));
assert!(!should_close_on_queue_full_streak(2)); assert!(!should_close_on_queue_full_streak(2, PressureState::Pressured));
assert!(should_close_on_queue_full_streak(3)); assert!(!should_close_on_queue_full_streak(3, PressureState::Pressured));
assert!(should_close_on_queue_full_streak(u8::MAX)); assert!(should_close_on_queue_full_streak(3, PressureState::Shedding));
assert!(should_close_on_queue_full_streak(
u8::MAX,
PressureState::Saturated
));
}
#[test]
fn fairness_retry_is_scheduled_only_when_queue_has_pending_bytes() {
let mut snapshot = WorkerFairnessSnapshot::default();
assert!(!should_schedule_fairness_retry(&snapshot));
snapshot.total_queued_bytes = 1;
assert!(should_schedule_fairness_retry(&snapshot));
}
#[test]
fn fairness_retry_delay_never_drops_below_one_millisecond() {
assert_eq!(fairness_retry_delay(0), Duration::from_millis(1));
assert_eq!(fairness_retry_delay(2), Duration::from_millis(2));
} }
#[test] #[test]

View File

@@ -37,20 +37,26 @@ pub(super) fn validate_proxy_secret_len(data_len: usize, max_len: usize) -> Resu
/// Fetch Telegram proxy-secret binary. /// Fetch Telegram proxy-secret binary.
#[allow(dead_code)] #[allow(dead_code)]
pub async fn fetch_proxy_secret(cache_path: Option<&str>, max_len: usize) -> Result<Vec<u8>> { pub async fn fetch_proxy_secret(
fetch_proxy_secret_with_upstream(cache_path, max_len, None).await cache_path: Option<&str>,
max_len: usize,
proxy_secret_url: Option<&str>,
) -> Result<Vec<u8>> {
fetch_proxy_secret_with_upstream(cache_path, max_len, proxy_secret_url, None).await
} }
/// Fetch Telegram proxy-secret binary, optionally through upstream routing. /// Fetch Telegram proxy-secret binary, optionally through upstream routing.
pub async fn fetch_proxy_secret_with_upstream( pub async fn fetch_proxy_secret_with_upstream(
cache_path: Option<&str>, cache_path: Option<&str>,
max_len: usize, max_len: usize,
proxy_secret_url: Option<&str>,
upstream: Option<Arc<UpstreamManager>>, upstream: Option<Arc<UpstreamManager>>,
) -> Result<Vec<u8>> { ) -> Result<Vec<u8>> {
let cache = cache_path.unwrap_or("proxy-secret"); let cache = cache_path.unwrap_or("proxy-secret");
// 1) Try fresh download first. // 1) Try fresh download first.
match download_proxy_secret_with_max_len_via_upstream(max_len, upstream).await { match download_proxy_secret_with_max_len_via_upstream(max_len, upstream, proxy_secret_url).await
{
Ok(data) => { Ok(data) => {
if let Err(e) = tokio::fs::write(cache, &data).await { if let Err(e) = tokio::fs::write(cache, &data).await {
warn!(error = %e, "Failed to cache proxy-secret (non-fatal)"); warn!(error = %e, "Failed to cache proxy-secret (non-fatal)");
@@ -91,14 +97,19 @@ pub async fn fetch_proxy_secret_with_upstream(
#[allow(dead_code)] #[allow(dead_code)]
pub async fn download_proxy_secret_with_max_len(max_len: usize) -> Result<Vec<u8>> { pub async fn download_proxy_secret_with_max_len(max_len: usize) -> Result<Vec<u8>> {
download_proxy_secret_with_max_len_via_upstream(max_len, None).await download_proxy_secret_with_max_len_via_upstream(max_len, None, None).await
} }
pub async fn download_proxy_secret_with_max_len_via_upstream( pub async fn download_proxy_secret_with_max_len_via_upstream(
max_len: usize, max_len: usize,
upstream: Option<Arc<UpstreamManager>>, upstream: Option<Arc<UpstreamManager>>,
proxy_secret_url: Option<&str>,
) -> Result<Vec<u8>> { ) -> Result<Vec<u8>> {
let resp = https_get("https://core.telegram.org/getProxySecret", upstream).await?; let resp = https_get(
proxy_secret_url.unwrap_or("https://core.telegram.org/getProxySecret"),
upstream,
)
.await?;
if !(200..=299).contains(&resp.status) { if !(200..=299).contains(&resp.status) {
return Err(ProxyError::Proxy(format!( return Err(ProxyError::Proxy(format!(

View File

@@ -2,6 +2,7 @@ use std::time::{Duration, Instant};
use bytes::Bytes; use bytes::Bytes;
use crate::protocol::constants::RPC_FLAG_QUICKACK;
use crate::transport::middle_proxy::fairness::{ use crate::transport::middle_proxy::fairness::{
AdmissionDecision, DispatchAction, DispatchFeedback, PressureState, SchedulerDecision, AdmissionDecision, DispatchAction, DispatchFeedback, PressureState, SchedulerDecision,
WorkerFairnessConfig, WorkerFairnessState, WorkerFairnessConfig, WorkerFairnessState,
@@ -114,6 +115,62 @@ fn fairness_keeps_fast_flow_progress_under_slow_neighbor() {
assert!(snapshot.total_queued_bytes <= 64 * 1024); assert!(snapshot.total_queued_bytes <= 64 * 1024);
} }
#[test]
fn fairness_prioritizes_quickack_flow_when_weights_enabled() {
let mut now = Instant::now();
let mut fairness = WorkerFairnessState::new(
WorkerFairnessConfig {
max_total_queued_bytes: 256 * 1024,
max_flow_queued_bytes: 64 * 1024,
base_quantum_bytes: 8 * 1024,
pressured_quantum_bytes: 8 * 1024,
penalized_quantum_bytes: 8 * 1024,
default_flow_weight: 1,
quickack_flow_weight: 4,
..WorkerFairnessConfig::default()
},
now,
);
for _ in 0..8 {
assert_eq!(
fairness.enqueue_data(10, RPC_FLAG_QUICKACK, enqueue_payload(16 * 1024), now),
AdmissionDecision::Admit
);
assert_eq!(
fairness.enqueue_data(20, 0, enqueue_payload(16 * 1024), now),
AdmissionDecision::Admit
);
}
let mut quickack_dispatched = 0u64;
let mut bulk_dispatched = 0u64;
for _ in 0..64 {
now += Duration::from_millis(1);
let SchedulerDecision::Dispatch(candidate) = fairness.next_decision(now) else {
break;
};
if candidate.frame.conn_id == 10 {
quickack_dispatched = quickack_dispatched.saturating_add(1);
} else if candidate.frame.conn_id == 20 {
bulk_dispatched = bulk_dispatched.saturating_add(1);
}
let _ = fairness.apply_dispatch_feedback(
candidate.frame.conn_id,
candidate,
DispatchFeedback::Routed,
now,
);
}
assert!(
quickack_dispatched > bulk_dispatched,
"quickack flow must receive higher dispatch rate with larger weight"
);
}
#[test] #[test]
fn fairness_pressure_hysteresis_prevents_instant_flapping() { fn fairness_pressure_hysteresis_prevents_instant_flapping() {
let mut now = Instant::now(); let mut now = Instant::now();
@@ -180,6 +237,12 @@ fn fairness_randomized_sequence_preserves_memory_bounds() {
} }
let snapshot = fairness.snapshot(); let snapshot = fairness.snapshot();
let (standing_recomputed, backpressured_recomputed) =
fairness.debug_recompute_flow_counters(now);
assert!(snapshot.total_queued_bytes <= 32 * 1024); assert!(snapshot.total_queued_bytes <= 32 * 1024);
assert_eq!(snapshot.standing_flows, standing_recomputed);
assert_eq!(snapshot.backpressured_flows, backpressured_recomputed);
assert!(fairness.debug_check_active_ring_consistency());
assert!(fairness.debug_max_deficit_bytes() <= 4 * 1024);
} }
} }