30 points by timeseries_dev 7 months ago flag hide 11 comments
johnsmill 7 months ago next
Fascinating project, I've been looking for a TSDB with built-in anomaly detection.
jane123 7 months ago next
Interesting! I'd like to know more about how the anomaly detection works.
mcsweeney 7 months ago prev next
I'd like to know what kind of hardware you used and if it's a resource hog?
johnsmill 7 months ago next
It's a pretty standard server setup and it performs quite well.
network_guru 7 months ago next
Standard setup might get tricky if you move to a cloud based architecture.
code_for_fun 7 months ago prev next
What are the limitations when scaling this up? In terms of storage and processing power?
optimus_spark 7 months ago next
Scaling up is dependent on the infrastructure choices and the chosen ML algorithms.
crypto_genius 7 months ago prev next
Why use a TSDB when ML algorithms can handle anomalies on their own?
big_data_dave 7 months ago next
ML algorithms usually need pre-processing and TSDBs provide the necessary temporal context.
engineer_by_heart 7 months ago next
Very true, it's helpful that this TSDB abstracts a lot of the pre-processing required for ML.
jane123 7 months ago prev next
Can't wait to implement this in our system, thanks for this great share!