Operations Retainer

Plenty of teams run Cassandra in production without a full-time specialist — the cluster mostly behaves, until it doesn't. The operations retainer puts a senior Cassandra engineer within reach by the hour, for the routine hygiene that prevents incidents and the escalations when one happens anyway.

Repair and compaction hygiene

Repairs that complete inside gc_grace_seconds, on a schedule that doesn't fight your peak traffic. Compaction strategies reviewed as workloads drift — the table that was fine on STCS two years ago may be a time-series table on the wrong strategy today. We keep both boring, which is the goal.

Monitoring and alerting

Dashboards and alerts on the metrics that predict trouble rather than announce it: pending compactions, tombstones scanned per read, partition-size growth, dropped mutations, hint accumulation, repair age per table. If your current monitoring only tells you when a node is down, this is the first month's work.

Incident escalation

When reads slow down or a node misbehaves, an engineer who has seen the failure mode before joins your bridge. We work the incident with your team — diagnosis, mitigation, and afterwards the write-up that keeps it from recurring.

Performance tuning

Heap and GC settings, compaction throughput, concurrent reads/writes, driver-side fixes, and the query-level work — because most "cluster is slow" reports trace back to a handful of queries and the partitions under them.

How it works

Hourly time and materials, sized to your needs month by month — some months that's a few hours of scheduled hygiene, some months it's an upgrade's worth of work. No minimum commitments dressed up as tiers; scope and staffing are agreed in an initial conversation and revisited whenever your situation changes.