Madu Mitha Ravi
Madu Mitha Ravi
@praveen-influx With the latest commit as of Feb 26th, - No more OOMs are observed 🎉 - 2 hours going good, whereas with the current code we OOM in about...
Point 1 - Complete and verified by Praveen, below are the details. - Ingestion rate of 120K rows per second across 50 tables - Approx. 27 GiB of data written...
### Query perf results on the branch (rebased version) Here is the [detailed document](https://docs.google.com/document/d/1_Sb2n2h7M1LEIOj_nkH-sTP30PttbE2wrYBkhLYjPqs/edit?usp=sharing) on the experiments and results. Overall, we are good to proceed with the merge and then...
I will do some quick runs and update in here. We can modify the issue based on evidence.
@hiltontj You suspicion is right. More outliers spike with the increase in concurrency.
@hiltontj Not a blocker, just a concern. We can take it up next week.
Working on the re-runs with the latest build (with fix), will update once I am done.
Results after the merge in on DF Threadpool. - Latencies have spiked compared to previous experiment - CPU usage has increased significantly , reaches 80% at concurrency of 4 (previously...
Trace Analysis of 4 concurrent threads, Measurement | Compare Item | Query Name | execute_stream_partitioned (ms) | sql_to_logical_plan (ms) | get_namespace (ms) | flight_planner (ms) | sql_to_physical_plan (ms) | query_planning...
Same query, but the tag value (device_id) differs. `SELECT * FROM last_cache('senec_data', 'ntag1-lv-1') where device_id in (....);`