Query concurrency limit change in InfluxDB 2.7 leads to higher latency for complex Flux queries
0 reputation · 10 Jun 2021, 06:14 UTC
Scenario
When the query-concurrency-limit setting is enabled in InfluxDB 2.7 to restrict simultaneous Flux query execution, users report increased latency for certain analytical queries that previously completed within sub‑second times.
Goal
Determine whether the observed latency increase stems from the query scheduler hitting the concurrency limit or from heightened storage I/O caused by the limited parallelism.
Uncertainty
It is unclear which internal metrics best reflect scheduler contention versus disk pressure, and how to isolate the effect without altering other configuration parameters.
Which specific counters in the _internal monitoring system indicate that the query scheduler is the bottleneck?
How can storage‑related metrics be distinguished from scheduler‑related metrics when concurrency is limited?
What reproducible workload characteristics (e.g., query shape, time range, cardinality) should be used to isolate the impact of the query-concurrency-limit setting?
influxdb