ReplicaLag

ReplicaLag is a web app (with optional lightweight desktop agent) that continuously measures and explains replication lag across distributed databases and streaming pipelines. Instead of just showing “lag = X seconds,” it correlates lag spikes with deploys, schema changes, compaction/GC events, network jitter, hotspot partitions, and cloud incidents. It ingests metrics and logs from common systems (Postgres logical replication, MySQL, MongoDB, Cassandra, Kafka, etc.) and produces a single timeline that answers: what changed, where, and what to do next. The MVP focuses on fast setup: connect via read-only credentials and exporters, auto-detect topology, and generate actionable alerts (“Replica in us-east-1 is 4m behind due to WAL retention + saturated IOPS”). This is a combination traditional + AI app: deterministic rules for known failure modes plus an AI layer to summarize incidents and propose next steps.