The JusDB Database Engineering Blog

Operate Kafka Connect database pipelines with current JDBC and Debezium settings, protected REST access, deliberate DLQ handling, and task-level monitoring.

Design DynamoDB access patterns with composite keys, sparse GSIs, paginated queries, transactions, and evidence-based capacity planning.

Manage Elasticsearch indexes with ILM policies, correct shard sizing, and force merge for historical indices. Covers hot/warm/delete phases and cluster health monitoring.

Logical replication in PostgreSQL has a critical gap: slots do not survive failover. PostgreSQL 17 finally fixes this with failover slots, plus adds row and column filtering for selective replication.

Optimize MongoDB aggregation pipelines by ordering stages correctly, using indexes in $match and $sort, projecting early, and tuning $lookup with pipeline filters.

Set up Debezium to stream MySQL and PostgreSQL changes to Kafka. Covers prerequisites, connector deployment, change event structure, and replication slot monitoring.

Use Redis Streams for event sourcing and message queues. Covers XADD, XREADGROUP, consumer groups, XAUTOCLAIM for stale messages, and Python producer/consumer examples.

MySQL Group Replication and Galera Cluster both promise multi-primary high availability, but they behave very differently under load, network partitions, and node failures.

Design Cassandra tables around query patterns, not entities. Covers partition key cardinality, hot partition prevention, time bucketing, and secondary indexes vs materialized views.

Connection exhaustion is one of the most common causes of database downtime at scale. Here is how production teams use PgBouncer and ProxySQL to handle 10,000+ concurrent connections.

ClickHouse Cloud is convenient but can cost 3-5x more than a well-tuned self-hosted cluster at scale. This breakdown shows exactly where the cost lines cross.

Diagnose and fix PostgreSQL query plan cache issues. Covers generic vs custom plans, plan_cache_mode setting, DEALLOCATE, statistics targets, and pg_stat_statements variance detection.