Tag
#performance
7 articles
LiveView Streams for Large Datasets: 10k Rows Without the Memory Bill
Phoenix LiveView streams solve a problem most dashboards pretend they don't have: the DOM keeps every row you ever rendered. Here's when to reach for streams, when to reach for pagination, and the memory math behind the decision.
ETS-Backed Registries: The Performance Trick Behind 552 Self-Registering Agents
Self-registration sounds like metaprogramming magic. It's actually just ETS + @after_compile + discipline. Here's exactly how Prismatic keeps 1,110 agent definitions (~70 runtime) and 128 OSINT tools discoverable in sub-millisecond time.
PostgreSQL and TimescaleDB Patterns for Intelligence Platforms
Database architecture patterns using Ecto schemas, TimescaleDB hypertables for time-series data, connection pooling strategies, and query optimization techniques.
LiveView Performance: Sub-250ms Page Loads
Achieving sub-250ms page loads in Phoenix LiveView requires careful attention to mount optimization, ETS caching, and error handling. This article shares the patterns we use across 30+ LiveView pages in production.
LiveView Performance Patterns for Complex Dashboards
Practical patterns for building high-performance Phoenix LiveView dashboards: defensive mounts, efficient assigns, PubSub-driven updates, and Chart.js integration with sub-150ms mount times.
ETS-Backed Registries: Sub-Millisecond Access in Elixir
ETS tables provide concurrent read access with sub-millisecond latency. The Prismatic Platform uses ETS for agent registries, OSINT tool catalogs, blog articles, and glossary terms with patterns you can apply to your own projects.
ETS-Backed Registries: Sub-Millisecond Data Access in Elixir
How Prismatic uses ETS tables as high-performance registries for agents, OSINT adapters, blog articles, and platform metrics. Patterns for compile-time loading, lazy initialization, and concurrent reads.