This message was deleted.
# supports
Slackbot
11/17/2022, 8:16 PMThis message was deleted.
a
Alexei Zenin
11/17/2022, 8:17 PMmetrics took a minute or 2 but traces still seems to keep going
Alexei Zenin
11/17/2022, 8:20 PMcurrently there are 18 million spans stored (~20 GB)
Alexei Zenin
11/17/2022, 8:25 PMseems it worked now. Took like 10 minutes, wonder if this can be optimized as it seems it scales with the amount of data
Alexei Zenin
11/17/2022, 8:25 PMcc @Prashant Shahi @Pranay
p
Pranay
11/18/2022, 12:38 AM
Yeah, sometimes it takes more time.
@Vishal Sharma may have more insights on this
v
Vishal Sharma
11/18/2022, 1:50 AM@Alexei Zenin Traces has around 5-6 tables and far more data than metrics. Setting TTL on clickhouse tables with large is very slow, that’s why we are using async api for setting TTL. For new users it is recommended to set TTL while data is less.
a
Alexei Zenin
11/22/2022, 4:39 PMI see.
Can this be optimized at all? I foresee a future where I will have TB of data in Clickhouse and will need to tweak the TTL. Will it take days or weeks to change at that point? I would assume I wouldn’t have to wipe out all data for the past month causing our team to lose data and start from scratch each time. Is this possible to optimize at all?
v
Vishal Sharma
11/23/2022, 5:53 AM@Alexei Zenin Clickhouse team is already working improving on optimizing TTL: https://github.com/ClickHouse/ClickHouse/issues/10128
Meanwhile we will search other ways to drop old data. You can track issue here: https://github.com/SigNoz/signoz/issues/1741
7 Views