does greptime buffer things by 10s at default? is there any way to query round trip from ingestion faster?
n
Ning
11/28/2025, 12:43 AM
greptimedb is a realtime database, anything you get an success response for ingestion can be queried. Could you provide more information about your 10s latency?
d
Daniel
11/28/2025, 7:29 PM
I’m using fluent-bit to send logs to greptime, it’s probably buffering there
Daniel
11/28/2025, 7:43 PM
ok yeah, it’s a fluent-bit issue
Daniel
11/28/2025, 7:44 PM
how does greptime handle real-time – does it keep some ingestion data in memory or does it do a full e2e cycle for any query. if you write in real-time does that mean its closing files on the remote s3 every second, or can it append-data?
n
Ning
11/29/2025, 12:06 AM
yes we have a memory buffer to accumulate data before flushing to s3. Our query engine will also treat this memory buffer as part of data source when processing query.
d
Daniel
11/29/2025, 2:02 AM
amazing, that’s so useful. are there docs on this mechanism?