Tuning Core OSS

1

Hey team, I’m trying to evaluate whether the v3 core solution is the right database for my product, but I’m seeing poor performance and I suspect I might be doing something wrong.

I work at a telecom company where we’re measuring 4G cells.

  • For cells, our tags are cell id and node id, with ~20 fields.

  • For UEs, we use ue id, cell id, and node id as tags, with ~25 fields.

  • Data precision is 1 second.

  • Most queries are aggregations over the last 10 seconds for specific cells or UEs.

The database is running on bare metal (HP DL360) with 32 CPUs and 32 GB memory. Queries are consistently taking over 1 second (not a network issue). Writes are asynchronous across N TCP sessions, but acknowledgments often take >500 ms.

What can I do to improve this? Or is InfluxDB maybe not the right solution for this kind of workload?

Thanks in advance!

2

Hi @DanShmirer InfluxDB 3 Core can certainly handle your requirements with ease, however few things to bear in mind:

  1. Your query time frame is within 72 hour window, if it’s beyond then suggest trying out InfluxDB 3 Enterprise (30 day free trial to try just with your email). Our engineers can also help in this case.
  2. You are leveraging caching - InfluxDB 3 Core & Ent have in memory cache built in such as ‘Last Value Cache’ & ‘Distint Value Cache’ that can significantly boost queryt performance (less than 10 ms). See Query the Latest Values in Under 10ms with the InfluxDB 3 Last Value Cache | InfluxData
  3. For writing, make sure you’re writing the data in batches. See Optimize writes to InfluxDB 3 Core | InfluxDB 3 Core Documentation

Happy to answer any follow up questions.

3

Hi Suyash, thanks for your reply! I’m still unclear on what’s causing the slow queries (>1s for 10s windows) and high write acks (>500 ms). Beyond batching, could you point me to specific schema or config issues (tags, shard setup, durability, etc.) that might explain this? Just trying to understand if this is tuning on my side or a Core limitation.

Thanks!

4

Sure, here you go: InfluxDB schema design recommendations and best practices | InfluxDB 3 Core Documentation Have a read through as schema design can affect the performance.