feat(kubenetes_logs): Annotate logs with node labels #12730
Conversation
✅ Deploy Preview for vector-project canceled.
|
github-actions
bot
added
domain: ci
This comment was marked as outdated.
This comment was marked as outdated.
Soak Test ResultsBaseline: 7fd73b5 ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. Changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%:
Fine details of change detection per experiment.
|
spencergilbert
left a comment
spencergilbert
left a comment
There was a problem hiding this comment.
Thanks for the contribution!
There was a problem hiding this comment.
Whoops - blocking for the moment:
NOTE!: Integrations tests do not work because access rights are required to list and watch node objects. When I faced this issue, I came up with an idea. What if we make nodes (and probably namespaces) annotators lazy-loaded? It will allow users to control each Vector pod's number of watch requests. Maybe this behavior is not explicit enough. I just wanted to share the idea.
We should update the helm chart to add nodes to the RBAC, and add this to the upgrade guide before merging in the change. Happy to do both of these if you'd like me to @nabokihms.
I'm interested in the lazy loading idea as well, any thoughts on how it'd work? (probably worth opening an issue regardless)
Soak regression likely resolved by merging in master
Soak Test ResultsBaseline: 50dd7a7 ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. No interesting changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%: Fine details of change detection per experiment.
|
Soak Test ResultsBaseline: 50dd7a7 ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. Changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%:
Fine details of change detection per experiment.
|
|
@spencergilbert Thank you for the review. I fixed imports, so everything should be fine now.
Ok, no problem. I will describe my idea in detail in a separate issue 👍
You are seducing me to drop everything on you 😅 If you don't mind, I'd like to do both myself to get familiar with the process (I hope it will help me with future contributions). |
😆 No problem! Upgrade guide for 0.22.0 can be found here, https://github.com/vectordotdev/vector/blob/master/website/content/en/highlights/2022-05-03-0-22-0-upgrade-guide.md, should be straightforward (0.21 for reference on the last time we added a new thing to RBAC) The Helm chart change (I haven't written a contributing doc, shame on me)... I'm a little foggy from vacation but I think it should just be bumping the chart version + running CI should catch anything else 🤔 |
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
Soak Test ResultsBaseline: 83bd797 ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. Changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%:
Fine details of change detection per experiment.
|
Soak Test ResultsBaseline: e20f40b ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. No interesting changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%: Fine details of change detection per experiment.
|
|
@nabokihms - should be able to get the integration tests passing locally now that the Helm chart has been updated, or pushing another commit here will run it in CI 👍 |
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Signed-off-by: m.nabokikh <maksim.nabokikh@flant.com>
Soak Test ResultsBaseline: ffc0f40 ExplanationA soak test is an integrated performance test for vector in a repeatable rig, with varying configuration for vector. What follows is a statistical summary of a brief vector run for each configuration across SHAs given above. The goal of these tests are to determine, quickly, if vector performance is changed and to what degree by a pull request. Where appropriate units are scaled per-core. The table below, if present, lists those experiments that have experienced a statistically significant change in their throughput performance between baseline and comparision SHAs, with 90.0% confidence OR have been detected as newly erratic. Negative values mean that baseline is faster, positive comparison. Results that do not exhibit more than a ±8.87% change in mean throughput are discarded. An experiment is erratic if its coefficient of variation is greater than 0.3. The abbreviated table will be omitted if no interesting changes are observed. No interesting changes in throughput with confidence ≥ 90.00% and absolute Δ mean >= ±8.87%: Fine details of change detection per experiment.
|
|
Thanks again @nabokihms! |
2 participants
This PR adds a new feature. With these changes, the vector agent will annotate all entries from kuberenetes_logs sources with node labels.
NOTE!: Integration tests do not work because access rights are required to list and watch node objects. When I faced this issue, I came up with an idea. What if we make nodes (and probably namespaces) annotators lazy-loaded? It will allow users to control each Vector pod's number of watch requests. Maybe this behavior is not explicit enough. I just wanted to share the idea.
Signed-off-by: m.nabokikh maksim.nabokikh@flant.com
Closes #7499