Cleared 3m ago · afreidah/s3-orchestrator #362 · Go · ★ 37 · MIT
Summary No trend metrics to predict when backends will fill. Operators must manually check usage dashboards and extrapolate. Impact Reactive scaling — learn about quota exhaustion from failed writes, not proactive alerts No "days until full" metric for capacity planning…
#363 · Go · ★ 37 · MIT · updated 7d ago
Summary The disaster recovery guide covers recovery procedures, but there are no decision-tree runbooks for on-call operators responding to alerts. A new operator seeing an alert for the first time has to read architecture docs to understand what to do. Needed Runbooks Top alert…
#787 · Go · ★ 34 · Apache-2.0 · updated 2mo ago
Problem The DS handler at workflow_handlers.go:1373 logs total_count and returned_count but does not log the filter parameters (component, environment, severity, priority). This made triaging #786 significantly harder — the hypothesis was that the DS SQL filtering was wrong, when…
#639 · Go · ★ 162 · Apache-2.0 · updated 2mo ago
Summary During rolling updates, each cluster's PerClusterPlacementStatus.ObservedResourceIndex is always set to the latest resource snapshot index, even when the binding for that cluster still points to an older resource snapshot. This means the per-cluster status does not…
#5086 · Go · ★ 2.1k · MIT · updated 3mo ago
Requirement Organizations running data warehouses on AWS Redshift need a first-class exporter that loads flag evaluation events directly into Redshift — without requiring an S3 intermediate stage. This enables BI teams to join feature flag data with application metrics already in…