New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prometheus OOM for lots of short time series #3780
Comments
Prometheus is not an event logging system, and attempting to use it that way will result in issues as you have discovered. I'd suggest looking at something like the ELK stack. |
Hello, Thank you for fast response. I understand that our time series is not continuous but I will like to use the Prometheus capabilities regarding time series rather than ELK stack. |
This isn't going to work out, Prometheus simply is not designed for this use case. Initial estimates indicate that you'd need at least 200GB of RAM to support this, and even then performance is likely to be poor. |
How it is calculated the 200GB of RAM? I am asking, so we can figure out if there is something else we could do to still use Prometheus. |
I just did a quick back of the envelope calculation. Prometheus is not intended for such high cardinalities of short lived data. |
This is expected behaviour. |
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
Hi all,
I have been playing with prometheus and trying to configure it so it will work.
What will be the optimal settings for this case? What caused the OOM issue?
What did you do?
We want to store around 50 million short time series (ts that are not appearing regularly) for a long period (lets say 5 years).
The number of time series is increasing exponentially, and we expect to reach a logarithmic slope at around 50 million time series.
The time series are pulled from a custom server written in NodeJS using siimon/prom-client npm library.
What did you expect to see?
Prometheus ingesting time series for at least a period of time until the storage is full.
What did you see instead? Under which circumstances?
Prometheus runs out of memory after 10 minutes. The docker is run with --restart always. After a few restarts, the docker appears to be up in docker ps -a, but docker stats shows no action happening. Also, I cannot longer enter the docker with exec -it <DOCKER_ID> /bin/sh (after a few restarts).
I have to mention that when making requests using the HTTP API, the response is Service Unavailable from the beginning.
Environment
- 24GB Ram
- 8 cores
- 500 GB HDD
- Docker version 17.09.1-ce, build 19e2cf6
- Starting prometheus docker with:
System information:
Linux 4.15.0-041500rc3-lowlatency x86_64
Prometheus version:
Prometheus configuration file:
The text was updated successfully, but these errors were encountered: