What happened
What this says
Current to 12 Oct 26Tariq reported slow dashboard performance when 8-10 users queried simultaneously despite fresh connector data. Bethan diagnosed the issue as warehouse spin-up delays when the warehouse enters standby.
Problem: Dashboard joining CRM and call records took 90 seconds to refresh under concurrent load, despite working connectors. Handling: Bethan reviewed query logs and identified execution time of 6-8 seconds with an additional 15-second warehouse spin-up, expected Snowflake behavior. She suggested the customer consider warehouse sizing or auto-suspend configuration options. Watch: This is a scaling issue, not a bug; resolution depends on customer infrastructure choices.
Details
Updated 9 Oct 26Trail
- created · Tariq Saleem 12 Jun 2025 09:47Hi, We've started to see slow load times on some of the dashboards when multiple users are querying at the same time, particularly the one that pulls data from the CRM and the call records together. It's not down, just sluggish. I've checked the connector load logs and both connectors have fresh data as of this morning. Can you help me work out where the bottleneck is? Cheers, Tariq
- public_reply · Bethan Lister 12 Jun 2025 11:16Hi Tariq, Just so I've got this right, the dashboard is loading slowly when there's concurrent traffic, even though the connectors are up to date. That points to a query performance issue on the Snowflake side rather than a data problem. Can you tell me: 1. How many records is the dashboard querying across? 2. How many concurrent users would you say are typical when you see the slowdown? 3. Are there any other dashboards or queries running at the same time, or just this one? That'll help me work out if it's the Snowflake warehouse size or the query itself. Thanks, Bethan
- public_reply · Tariq Saleem 13 Jun 2025 10:11The dashboard is pulling about 200 licensed users of activity data joined with trade account details, roughly two years' worth. Slowdown happens when we've got eight to ten people running reports at the same time. This morning I ran a test and it took about ninety seconds to refresh. Normally it's instant. I haven't changed anything on our side.
- public_reply · Bethan Lister 16 Jun 2025 10:58Thanks for that detail. Ninety seconds for a dashboard that big under load suggests the Snowflake warehouse might be spinning down between queries. I'm going to look at the query logs on our end to check the execution times and see if there's a tuning option.
- internal_note · Bethan Lister 16 Jun 2025 16:11Reviewed the query logs. Execution time is six to eight seconds per query, so the warehouse is responding fine. The ninety-second total includes a fifteen-second warehouse spin-up time each time, which means it's going into standby between user queries. This is expected Snowflake behaviour. Options: warehouse size is fine as is, or we can talk about auto-suspend settings. Will loop Felix in, this may be a Snowflake tuning question rather than our code.
- status_change · 22 Jun 2025 23:59