Set `storage.local.target-heap-size` in Prometheus 2.0.0-rc3

[Marius Mitrofan]

There was `-storage.local.target-heap-size=#{prometheus_target_heap_size} \` that I was computing in my Chef recipes in version 1.* but apparently this flag doesn't exist anymore.

So.. what now? I'm asking this because it seems that my heap size doesn't go above ~ 2.5 GB anymore (the same problem I had before setting up the flag in version 1.*).

[Ben Kochie]

Prometheus 2.0 has a completely different TSDB and a completely different way to handle caching series data in memory.  Most of the target heap size flag was used for was to control block caching.

This is now entirely handled by kernel page cache, so the whole concept is obsolete.

On Mon, Nov 6, 2017 at 11:23 AM, 'Marius Mitrofan' via Prometheus Users <promethe...@googlegroups.com> wrote:

There was `-storage.local.target-heap-size=#{prometheus_target_heap_size} \` that I was computing in my Chef recipes in version 1.* but apparently this flag doesn't exist anymore.

So.. what now? I'm asking this because it seems that my heap size doesn't go above ~ 2.5 GB anymore (the same problem I had before setting up the flag in version 1.*).

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-users+unsubscribe@googlegroups.com.
To post to this group, send email to prometheus-users@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/7da3e64f-55d6-4641-a35d-686426629f6b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Marius Mitrofan]

Awesome, thanks!

On Monday, November 6, 2017 at 12:33:51 PM UTC+2, Ben Kochie wrote:

Prometheus 2.0 has a completely different TSDB and a completely different way to handle caching series data in memory.  Most of the target heap size flag was used for was to control block caching.

This is now entirely handled by kernel page cache, so the whole concept is obsolete.

On Mon, Nov 6, 2017 at 11:23 AM, 'Marius Mitrofan' via Prometheus Users <promethe...@googlegroups.com> wrote:

There was `-storage.local.target-heap-size=#{prometheus_target_heap_size} \` that I was computing in my Chef recipes in version 1.* but apparently this flag doesn't exist anymore.

So.. what now? I'm asking this because it seems that my heap size doesn't go above ~ 2.5 GB anymore (the same problem I had before setting up the flag in version 1.*).

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
To post to this group, send email to promethe...@googlegroups.com.

[David Karlsen]

I run prometheus in a docker-container. Will this affect how I should configure things/define resource-constraints for the container?

mandag 6. november 2017 11.33.51 UTC+1 skrev Ben Kochie følgende:

Prometheus 2.0 has a completely different TSDB and a completely different way to handle caching series data in memory.  Most of the target heap size flag was used for was to control block caching.

This is now entirely handled by kernel page cache, so the whole concept is obsolete.

On Mon, Nov 6, 2017 at 11:23 AM, 'Marius Mitrofan' via Prometheus Users <promethe...@googlegroups.com> wrote:

There was `-storage.local.target-heap-size=#{prometheus_target_heap_size} \` that I was computing in my Chef recipes in version 1.* but apparently this flag doesn't exist anymore.

So.. what now? I'm asking this because it seems that my heap size doesn't go above ~ 2.5 GB anymore (the same problem I had before setting up the flag in version 1.*).

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
To post to this group, send email to promethe...@googlegroups.com.

[Ben Kochie]

I'm not very familiar with how Docker accounts page cache to containers, and controls memory limits for other containers.

In theory, you will want to size your memory limits such that you have memory reserved for both the Prometheus process and some amount for caching metrics depending on your query needs.

To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-users+unsubscribe@googlegroups.com.
To post to this group, send email to prometheus-users@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/a5a45f9c-fddd-40f0-bac9-9aa1064d5444%40googlegroups.com.