ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
Learn how to use Apache Pulsar tenants and namespaces to isolate production and staging environments, set namespace‑level policies, and mitigate noisy‑neighbor risks.
Shared subscriptions in Pulsar destroy per-key ordering by round-robining messages, while failover subscriptions refuse to scale. Key_Shared restores ordering per key by hashing keys onto consumers—here is how it works, where it breaks, and how to verify it.
A technical guide on choosing between Exclusive, Failover, Shared, and Key_Shared subscription modes in Apache Pulsar to balance message ordering and throughput.
Pulsar's built-in schema registry enforces compatibility at the broker, turning schema evolution from a production incident into a controlled operation. Here's how the policies work and where they bite.
If your Pulsar cluster is running out of disk space, tiered storage can move old messages to S3 without changing consumer APIs. Learn the architecture, how to set it up, and the trade‑offs you’ll face.
Pulsar's Key_Shared subscription preserves per-key ordering while scaling consumers horizontally. Here's how it routes messages, a working example, and the hot-key and rebalancing trade-offs to design around.
How to offload older Pulsar messages to object storage using namespace‑level offload thresholds and deletion lag, with verification steps and rollback options.
Pulsar's Key_Shared subscription keeps every message key in order while the consumer group scales. How the hash routing works, what hot keys and rebalances cost, and how to verify it on your cluster.
Pulsar's tiered storage moves cold segments from BookKeeper to S3 or GCS automatically. Here's how offload triggers, how to configure and verify it, and the geo-replication and GC trade-offs to plan for.
pulsar