ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Databases & messaging · #cassandra
Use Cassandra lightweight transactions with IF NOT EXISTS and SERIAL consistency to prevent duplicate inserts on a single partition.
Decide which Jaeger backend—Cassandra, Elasticsearch, or PostgreSQL—fits your scale and operational needs. This guide compares write/read paths, scalability, and pagination, then walks through a concrete Cassandra deployment and validation steps.
TimeWindowCompactionStrategy lets Cassandra drop expired SSTables without rewriting them, but only when partition buckets, TTL, and window settings align. Here is how to design and verify that fit.
A technical guide for selecting a Jaeger storage backend, comparing the trade-offs between in-memory, Elasticsearch, and Cassandra for production tracing.
Learn how Cassandra Lightweight Transactions use Paxos to provide linearizable compare-and-set on a single partition, with a concrete UPDATE example, latency numbers, common pitfalls, and verification steps you can run on a 3‑node cluster.
Compare Cassandra consistency levels like ONE, QUORUM, and ALL to balance latency, durability, and availability, with a concrete validation workflow.
Stop using ALLOW FILTERING. Learn how Cassandra Materialized Views allow you to re-index data for fast reads while shifting synchronization logic from your app to the database.
Lightweight Transactions in Cassandra provide compare-and-set semantics for a single partition, but they carry coordination overhead. Learn when to use them, what boundaries they enforce, and how to measure their impact.
Read latency spikes on specific Cassandra tables with normal writes often signal tombstone accumulation from deletes and TTLs. This guide covers recognition, cause diagnosis, ordered checks with nodetool and logs, and fixes tied to findings with verification steps.
jaegertracing
elasticsearch