Best Approach for Real-Time Search Index Updates with Manticore ??

Hi everyone,

I’ve been working with Manticore Search for a few weeks now and really love the flexibility it offers—especially the full-text capabilities. However, I’m running into a challenge I hope someone here can help with.

My use case involves a high-frequency data stream (user-generated content) where I need real-time or near real-time indexing. I’m currently using the JSON over HTTP interface for inserts, but I’m noticing some delays and inconsistencies when querying right after an update.

My questions:

What’s the most reliable method for real-time inserts? Should I switch to using insert via MySQL protocol or continue with HTTP?

Are there specific config optimizations (e.g., rt_flush_period or rt_mem_limit) that can improve performance for high-frequency inserts?

Any best practices for minimizing indexing latency in production?

Appreciate any guidance from those who’ve scaled Manticore in similar scenarios! I have gone through these resources Best approach to indexing and searching relational data salesforce course in bangalore but still looking for more help.

Thanks in advance!

Derek

@dereektheler Hi! Just to confirm - are you using the /update endpoint, getting an OK response, but then when you fetch the document afterward, the value doesn’t seem to be updated? If so, that might be a bug. Could you help us reproduce it with a minimal example?