#community-help

Discussion on Managing Slow Request Logs on Typesense Cloud

TLDR Ethan inquired about managing --log-slow-requests-time-ms server parameter for a cloud cluster. Jason stated these measurements are managed on their side for diagnostic purposes and provided recent query performance details.

Powered by Struct AI
eyes1
9
1mo
Solved
Join the chat
Aug 12, 2023 (1 month ago)
Ethan
Photo of md5-43493fa2e131ee4b59d15256c7c083bf
Ethan
12:36 AM
We want to log slow requests and found this server parameter --log-slow-requests-time-ms but noticed its managed separately in typesense cloud - how can we toggle this parameter for a cloud cluster?
Image 1 for We want to log slow requests and found this server parameter <code>--log-slow-requests-time-ms</code>  but noticed its managed separately in typesense cloud - how can we toggle this parameter for a cloud cluster?
Jason
Photo of md5-8813087cccc512313602b6d9f9ece19f
Jason
12:38 AM
We only enable it on our side for diagnostic purposes on a need basis when we need to debug some reported issues
12:39
Jason
12:39 AM
Are you seeing particular slow query issues?
Ethan
Photo of md5-43493fa2e131ee4b59d15256c7c083bf
Ethan
12:40 AM
Rushil mentioned occasionally seeing some slow initial queries for company_id facet so we wanted to add a log for requests taking > 2s to monitor
Jason
Photo of md5-8813087cccc512313602b6d9f9ece19f
Jason
12:42 AM
I don’t see any search latency spikes just looking at historical metrics, in the last 24 hours…
12:43
Jason
12:43 AM
Avg is at ~15ms and max is at 200ms
12:43
Jason
12:43 AM
These same measurements are what Typesense uses to log slow queries
Ethan
Photo of md5-43493fa2e131ee4b59d15256c7c083bf
Ethan
12:43 AM
ah I see - Wednesday was when he mentioned the slow requests
Jason
Photo of md5-8813087cccc512313602b6d9f9ece19f
Jason
12:45 AM
Last 3 days max latency is 350ms, avg 25ms
eyes1