From 47d49595885845f1005ed90509351c4c6d7e83ce Mon Sep 17 00:00:00 2001 From: jgrusewski Date: Mon, 9 Mar 2026 20:53:36 +0100 Subject: [PATCH] fix(train): correct --output to --output-dir arg, add Prometheus annotations The train_baseline_rl binary expects --output-dir, not --output. Also adds prometheus.io scrape annotations to job pod templates so training metrics appear in Grafana. Co-Authored-By: Claude Opus 4.6 --- infra/scripts/train.sh | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/infra/scripts/train.sh b/infra/scripts/train.sh index 0a2f65bda..7f8141572 100755 --- a/infra/scripts/train.sh +++ b/infra/scripts/train.sh @@ -157,7 +157,7 @@ generate_job_manifest() { local cli_args=("--model" "$model" "--symbol" "$SYMBOL" "--data-dir" "/data/futures-baseline") cli_args+=("--mbp10-data-dir" "/data/futures-baseline-mbp10") cli_args+=("--trades-data-dir" "/data/futures-baseline-trades") - cli_args+=("--output" "/output") + cli_args+=("--output-dir" "/output") if [[ "$PRESET" == "hyperopt" ]]; then cli_args+=("--trials" "$TRIALS") fi @@ -186,6 +186,10 @@ spec: ttlSecondsAfterFinished: 600 template: metadata: + annotations: + prometheus.io/scrape: "true" + prometheus.io/port: "9094" + prometheus.io/path: "/metrics" labels: app.kubernetes.io/part-of: foxhunt app.kubernetes.io/component: training-workflow