Training jobs are short-lived K8s Jobs that can't be reliably scraped
by Prometheus. This adds a TrainingMetricsPusher that pushes epoch-level
metrics (loss, accuracy, throughput, gradient health, checkpoints) to
the Pushgateway so they persist for Prometheus to scrape.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a reusable GrpcMetricsLayer that can be applied to any tonic server
via Server::builder().layer() to automatically emit three Prometheus
metric families for every gRPC handler: started_total, handled_total
(with status code), and handling_seconds histogram.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds a dedicated Prometheus metrics endpoint on port 9098 (configurable
via METRICS_PORT env var) with counters for HTTP requests, histograms for
request duration, gauges for uptime and active WebSocket connections, and
counters for WebSocket messages. Metrics are registered using once_cell
Lazy statics with abort-on-failure to satisfy the crate's deny(unwrap)
lint policy.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The #[instrument] import was incorrectly inserted inside a `use super {}`
block by the previous commit, causing syntax errors.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace direct tracing_subscriber::fmt::init() calls with the unified
common::observability::init_observability() in api_gateway, web-gateway,
data_acquisition_service, broker_gateway_service, and trading_agent_service.
All 8 services now emit JSON logs + OTLP traces to Tempo.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Previously init_observability() called init_json_logger() (which set the
global subscriber with a fmt layer) and then init_tracing() (which only
set the global tracer provider). Because no tracing_opentelemetry::layer()
was ever added to the subscriber, #[instrument] spans never reached Tempo.
Now init_observability() builds a single Registry subscriber that
composes an EnvFilter, a JSON fmt layer, and an optional OTLP layer
(via tracing_opentelemetry). The OTLP layer gracefully degrades to None
when Tempo is unavailable, so logging always works.
Also adds build_otel_tracer() to tracing_config.rs which returns
Option<Tracer> for the caller to wrap in an OpenTelemetryLayer inside
their subscriber stack, letting the S type parameter be inferred
correctly by the compiler.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Web-gateway routes through api-gateway for the foxhunt.tli proto, but
api-gateway requires JWT Bearer auth on all gRPC requests. This adds a
service_auth module that mints short-lived service JWTs (matching the
api-gateway's expected claims: iss, aud, roles, permissions) and injects
them into all 15 REST proxy calls and 4 gRPC stream bridges.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Web-gateway routing:
- Point TRADING_SERVICE_URL at api-gateway (proto mismatch fix)
Web-gateway uses foxhunt.tli.TradingService proto but was connecting
directly to trading-service which implements trading.TradingService.
api-gateway already proxies Subscribe* → Stream* correctly.
GitLab KAS:
- Disable gitlab_kas in appConfig to stop sidekiq NotifyGitPushWorker
errors (KAS pod was already disabled but Rails still tried to connect)
Trading service monitoring (3 stubs → real):
- AcknowledgeAlert: real alert lookup + state mutation in shared store
- GetActiveAlerts: returns actual active alerts from in-memory store
- StreamAlerts: now persists generated alerts (capped at 1000 entries)
Trading service ML streams (2 stubs → real):
- StreamModelMetrics: emits real inference_count, error_count, latency
per model every N seconds from the RuntimeModelInfo registry
- StreamSignalStrength: emits per-symbol signal aggregation from model
ensemble weights and latency confidence
Backtesting service:
- stop_backtest: real CancellationToken cancellation (was no-op)
Tokens stored per-backtest, execute_backtest wraps strategy call
in tokio::select! for immediate cancellation
Deleted 7 empty placeholder files:
- 4 Wave D regime stubs (dynamic_stops, ensemble, performance_tracker,
position_sizer) — comment-only files, never wired
- 2 Wave 3 feature stubs (microstructure, statistical)
- 1 PPO stub (unified_ppo.rs — empty struct definitions)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add ohlcv_bars table migration (050) for trading_service prediction loop
- Fix kube-state-metrics RBAC: add admissionregistration.k8s.io and
certificates.k8s.io API groups to suppress forbidden errors
- Fix web-gateway gRPC stream reconnect: detect Unimplemented via both
status code and error message text (proxy may remap codes)
- Add automated migration step to CI deploy pipeline with tracking table
(_applied_migrations) so migrations are applied before service rollouts
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Disable file logging in init_observability — use console-only (stdout).
Containerd captures stdout, Promtail ships to Loki. Eliminates
"Failed to create log directory: Permission denied" errors on 3 services.
- Remove LOG_DIR env vars and /app/logs emptyDir volumes from 5 deployments.
- Add OTEL_EXPORTER_OTLP_ENDPOINT=tempo to 5 services that were missing it
(api-gateway, web-gateway, data-acquisition, trading-agent, broker-gateway).
- Make OTLP tracing init non-fatal — services degrade gracefully if Tempo is
unavailable instead of failing entire observability stack.
- Stop web-gateway gRPC stream retry on Unimplemented status. When trading-service
doesn't implement streaming endpoints, log once and stop instead of retrying
every 60s indefinitely.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove artificial swarm_size cap from plan_hyperopt() that limited
concurrent trials to n_particles (default 20). Now concurrency is
purely VRAM-driven with a hardware cap of 128 threads.
optimize_parallel() auto-scales n_particles to match GPU budget:
- L4 24GB: ~65 concurrent DQN trials (was 20)
- H100 80GB: 128 concurrent DQN trials (was 20)
- CPU/small GPU: falls back to configured n_particles
max_trials scales proportionally to ensure 3+ PSO iterations
for convergence with larger swarms.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Auto-detect now considers both CPU count and GPU VRAM when choosing
concurrent trial count. Prevents OOM on smaller GPUs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
optimize_parallel() now uses HardwareBudget::plan_hyperopt() to compute
max concurrent trials based on model VRAM footprint. Includes ramp-up
(start at half concurrency), VRAM watchdog (check free memory before
spawning), and graceful degradation to sequential.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Computes optimal (concurrency, batch_size_bounds) based on model memory
footprint vs detected GPU VRAM. Scales from CPU-only (sequential) through
H100 80GB (full swarm parallel). 20% safety margin for CUDA allocator.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
PSO now controls batch_size within [64, 4096] range. HardwareBudget
adjusts upper bound based on detected VRAM. The .min(160.0) clamp was
silently overriding PSO's choices on any GPU larger than 3050 Ti.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- init_observability now reads LOG_DIR env var for log directory (falls back
to relative logs/ path). Gracefully disables file logging if directory
cannot be created instead of failing the entire observability init.
- Add ServiceAccount, Role, and RoleBinding for ml-training-service to
create/get/list/watch/delete batch/v1 Jobs (K8s training dispatch).
- Add LOG_DIR=/app/logs env to ml-training-service, trading-service, and
backtesting-service deployment YAMLs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Single model_memory_mb field now serves as the sole source of truth for
base model memory. Removes the redundant base_model_memory_mb field,
legacy dual-path branches in calculate_optimal_batch_size() and
max_safe_batch_size(), and the associated legacy test.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
AutoBatchSizer was computing an "optimal" batch_size (128 on L4) and
overriding the PSO-chosen value. On an L4 24GB this meant 3.5% VRAM
utilization per trial. Now the trainer only clamps if the batch_size
would actually OOM, letting PSO explore the full [64, 4096] range.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The GPU experience collector only supported DQNAgentType::Standard, but
enable_regime_qnetwork defaults to true, creating RegimeConditional
agents. This silently fell back to CPU with a debug! message invisible
at INFO log level.
- Add primary_head() accessor to RegimeConditionalDQN (returns trending head)
- Extract weights from primary head for both collector init and weight sync
- Upgrade debug! to warn! for missing GPU collection prerequisites
- Add warn! to PPO for silent GPU skip when raw market data not uploaded
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
These files were part of the Phase 2 cuda_pipeline but were untracked
and not included in previous commits. Required for compilation with
--features ml/cuda.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Phase 3 of the CUDA pipeline: both trainers now take a GPU-first path
for experience collection (128×500 = 64K experiences per kernel launch,
zero CPU-GPU roundtrips per timestep) with automatic CPU fallback.
- DQN: upload features alongside targets, GPU collection branch before
CPU loop, gpu_batch_to_experiences() conversion into replay buffer
- PPO: set_raw_market_data() for CudaSlice upload, GPU collection
branch bypasses collect_rollouts + prepare_training_batch entirely,
gpu_batch_to_trajectory_batch() conversion with in-kernel GAE
- Configurable GPU batch sizes (gpu_n_episodes, gpu_timesteps_per_episode)
and trading params (initial_capital, avg_spread) via hyperparameters
- SAFETY training diagnostics downgraded from warn! to debug!
- 4 new batch index-math validation tests in cuda_pipeline
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Move 10 shared device functions (gpu_random, leaky_relu, matvec_leaky_relu,
action_to_exposure, action_to_tx_cost, barrier_init/check/reset,
diversity_entropy, curiosity_inference) and shared constants from
dqn_experience_kernel.cu into a new common_device_functions.cuh header.
The DQN kernel now expects the common header to be prepended via NVRTC
source concatenation at compile time. DQN-specific functions
(q_forward_dueling, argmax_q) remain in the kernel file. This enables
the upcoming PPO kernel to reuse the same shared functions.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds GpuExperienceCollector field to DQNTrainer, initializes it when
dueling networks and curiosity module are present on CUDA device, and
syncs GPU weight copies after each training epoch.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Three fixes from spec review:
1. diversity_entropy returns -0.1 penalty (threshold < 1.0) instead of raw entropy
2. barrier_done included in episode termination (barrier hit ends episode)
3. step_in_episode counter resets properly on mid-loop episode reset
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace raw f64 price/quantity in SmallBatchOptimizer with typed
financial values. Conversion to f64 happens at the SIMD boundary
(_mm256_loadu_pd) only. Delete 3 orphaned dead-code files:
financial_safe.rs, simd_optimizations.rs, tests/financial_tests.rs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Aliased re-exports avoid collision with existing types::Price/Quantity/Money
during migration. Will rename to canonical names in Phase 6 cleanup.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Ratio arithmetic (Add, Sub, Mul, Div) now clamps overflow to
f64::MAX/f64::MIN instead of producing Inf from finite inputs
- Cross-type Price*Quantity→Money and Money/Quantity→Price use
saturating i128→i64 conversion via clamp instead of silent truncation
- Added clamp_finite() and saturating_i128_to_i64() helper functions
- Added 9 edge-case tests: 89/89 passing, clippy clean
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add Serialize/Deserialize to all four financial types
- Ratio::ln returns 0.0 for non-positive inputs (prevents NaN leak)
- Ratio::exp clamps to f64::MAX on overflow (prevents Inf leak)
- Ratio::powi clamps to f64::MAX/MIN on overflow
- Add 4 invariant-preservation tests (80 total)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace all `.to_string().parse::<f32/f64>()` patterns with
`num_traits::ToPrimitive` methods (`.to_f32()`, `.to_f64()`).
Each string roundtrip heap-allocated per conversion — fatal in
DQN hot loop (300K+ bars × epochs). Decimal stays as canonical
financial type; conversions happen at GPU/float boundaries only.
Also fixes blocking_read() in async context (risk_integration.rs).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
P2-C reversal/cash warns fire thousands of times per hyperopt trial
during backtest simulation. Also removes stale Kelly patch file (already
applied).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
These fire every training step when portfolio isn't populated yet —
extremely noisy during hyperopt with parallel trials.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
