feat: Update retention and concurrency for Thanos #461
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
schwesig
requested review from
larsks,
computate,
harshil-codes,
dystewart and
jbasu01
larsks
reviewed
Jul 1, 2024
schwesig
force-pushed
the
20240628_thanos_retention
branch
from
5324b7a
to
8934026
Compare
@larsks, thanks for the feedback. I was too quick and dirty on this, and it was too clear in my head, not so much for the uninvolved reader, though :-) |
computate
approved these changes
Jul 1, 2024
This PR addresses the retention rate issues as discussed in nerc-project/operations#618 (comment) (having more than 30d raw etc.). The changes include updating the retention and concurrency settings for the Thanos Compactor to enhance observability and metrics performance. We will stay with the defaults where possible, adding remarks with the defaults to better understand the next changes or possible errors. Changes to focus on the needs for class, cost, and invoice analysis, as for future predictions: - Updated `retentionResolutionRaw` from 30d to 90d (quarterly high details for deep analysis, especially GPUs) - Updated `retentionResolution5m` from 90d to 360d (for cost, usage, and invoices; 15 minutes could be enough, but is not a default option) - Set `retentionResolution1h` to 0d (retain forever, following the default and recommendation) - Added `blockDuration`, `cleanupInterval`, `deleteDelay`, `retentionInLocal`, `consistencyDelay`, `compactConcurrency`, and `downsampleConcurrency` settings: even if staying in the default, making the options visible in case of possible future changes) These changes aim to optimize data retention & resolution for needed use cases and ensure better performance. References: 1. [Thanos Compact Component](https://thanos.io/tip/components/compact.md/) 2. [Recommendations for Running Thanos and Prometheus](https://zapier.com/blog/five-recommendations-when-running-thanos-and-prometheus/) 3. [Red Hat Advanced Cluster Management Observability](https://docs.redhat.com/en/documentation/red_hat_advanced_cluster_management_for_kubernetes/2.9/html/observability/customizing-observability#adding-advanced-config:~:text=is%20not%20displayed.-,4.3.%C2%A0Adding%20advanced%20configuration%20for%20retention,-Add%20the%20advanced) Signed-off-by: /Thor(sten)?/ Schwesig <89909507+schwesig@users.noreply.github.com>
larsks
force-pushed
the
20240628_thanos_retention
branch
from
a494d61
to
02436f5
Compare
|
I rebased this on |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR addresses the retention rate issues as discussed in nerc-project/operations#618 (comment) (having more than 30d raw etc.).
The changes include updating the retention and concurrency settings for the Thanos Compactor to enhance observability and metrics performance.
We will stay with the defaults where possible, adding remarks with the defaults to better understand the next changes or possible errors.
Changes to focus on the needs for class, cost, and invoice analysis, as for future predictions:
retentionResolutionRawfrom 30d to 90d (quarterly high details for deep analysis, especially GPUs)retentionResolution5mfrom 90d to 360d (for cost, usage, and invoices; 15 minutes could be enough, but is not a default option)retentionResolution1hto 0d (retain forever, following the default and recommendation)blockDuration,cleanupInterval,deleteDelay,retentionInLocal,consistencyDelay,compactConcurrency, anddownsampleConcurrencysettings: even if staying in the default, making the options visible in case of possible future changes)These changes aim to optimize data retention & resolution for needed use cases and ensure better performance.
References: