semantic-conventions/docs/non-normative/k8s-migration.md

256 lines
15 KiB
Markdown

<!--- Hugo front matter used to generate the website version of this page:
linkTitle: K8s migration
--->
# K8s semantic convention stability migration
Due to the significant number of modifications and the extensive user base
affected by them, existing K8s instrumentations published by
OpenTelemetry are required to implement a migration plan that will assist users in
transitioning to the stable K8s semantic conventions.
When existing K8s instrumentations published by OpenTelemetry are
updated to the stable K8s semantic conventions, they:
- SHOULD introduce an environment variable `OTEL_SEMCONV_STABILITY_OPT_IN` in
their existing major version, which accepts:
- `k8s` - emit the stable k8s conventions, and stop emitting
the old k8s conventions that the instrumentation emitted previously.
- `k8s/dup` - emit both the old and the stable k8s conventions,
allowing for a phased rollout of the stable semantic conventions.
- The default behavior (in the absence of one of these values) is to continue
emitting whatever version of the old k8s conventions the
instrumentation was emitting previously.
- Need to maintain (security patching at a minimum) their existing major version
for at least six months after it starts emitting both sets of conventions.
- May drop the environment variable in their next major version and emit only
the stable k8s conventions.
Specifically for the Opentelemetry Collector:
The transition will happen through two different feature gates.
One for enabling the new schema called `semconv.k8s.enableStable`,
and one for disabling the old schema called `semconv.k8s.disableLegacy`. Then:
- On alpha the old schema is enabled by default (`semconv.k8s.disableLegacy` defaults to false),
while the new schema is disabled by default (`semconv.k8s.enableStable` defaults to false).
- On beta/stable the old schema is disabled by default (`semconv.k8s.disableLegacy` defaults to true),
while the new is enabled by default (`semconv.k8s.enableStable` defaults to true).
- It is an error to disable both schemas
- Both schemas can be enabled with `--feature-gates=-semconv.k8s.disableLegacy,+semconv.k8s.enableStable`.
<!-- toc -->
- [Summary of changes](#summary-of-changes)
- [K8s network metrics](#k8s-network-metrics)
- [K8s Deployment metrics](#k8s-deployment-metrics)
- [K8s ReplicaSet metrics](#k8s-replicaset-metrics)
- [K8s ReplicationController metrics](#k8s-replicationcontroller-metrics)
- [K8s StatefulsSet metrics](#k8s-statefulsset-metrics)
- [K8s HorizontalPodAutoscaler metrics](#k8s-horizontalpodautoscaler-metrics)
- [K8s DaemonSet metrics](#k8s-daemonset-metrics)
- [K8s Job metrics](#k8s-job-metrics)
- [K8s Cronjob metrics](#k8s-cronjob-metrics)
- [K8s Namespace metrics](#k8s-namespace-metrics)
<!-- tocstop -->
## Summary of changes
This section summarizes the changes made to the K8s semantic conventions
from a range of versions. Each starting version shows all the changes required
to bring the conventions to stable (TODO: link to specific version once it exists).
### K8s network metrics
The K8s network metrics implemented by the Collector and specifically the
[kubeletstats](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.112.0/receiver/kubeletstatsreceiver/documentation.md)
receiver were introduced as semantic conventions
in [v1.29.0](https://github.com/open-telemetry/semantic-conventions/blob/v1.29.0/docs/system/k8s-metrics.md).
The changes in their attributes are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|--------------------------|
| `interface` | `network.interface.name` |
| `direction` | `network.io.direction` |
<!-- prettier-ignore-end -->
### K8s Deployment metrics
The K8s Deployment metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1636](https://github.com/open-telemetry/semantic-conventions/pull/1636) (TODO: replace with SemConv version once
available).
The changes in their metric names and types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|--------------------------------------------------------------|
| `k8s.deployment.desired` (type: `gauge`) | `k8s.deployment.desired_pods` (type: `updowncounter`) |
| `k8s.deployment.available` (type: `gauge`) | `k8s.deployment.available_pods` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s ReplicaSet metrics
The K8s ReplicaSet metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1636](https://github.com/open-telemetry/semantic-conventions/pull/1636) (TODO: replace with SemConv version once
available).
The changes in their metric names and types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|---------------------------------------------------------|
| `k8s.replicaset.desired` (type: `gauge`) | `k8s.replicaset.desired_pods` (type: `updowncounter`) |
| `k8s.replicaset.available` (type: `gauge`) | `k8s.replicaset.available_pods` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s ReplicationController metrics
The K8s ReplicationController metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1636](https://github.com/open-telemetry/semantic-conventions/pull/1636) (TODO: replace with SemConv version once
available).
The changes in their metric names and types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|--------------------------------------------------------------------|
| `k8s.replication_controller.desired` (type: `gauge`) | `k8s.replicationcontroller.desired_pods` (type: `updowncounter`) |
| `k8s.replication_controller.available` (type: `gauge`) | `k8s.replicationcontroller.available_pods` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s StatefulsSet metrics
The K8s StatefulsSet metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1636](https://github.com/open-telemetry/semantic-conventions/pull/1636) (TODO: replace with SemConv version once
available).
The changes in their metric types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|----------------------------------------------------------|
| `k8s.statefulset.desired_pods` (type: `gauge`) | `k8s.statefulset.desired_pods` (type: `updowncounter`) |
| `k8s.statefulset.ready_pods` (type: `gauge`) | `k8s.statefulset.ready_pods` (type: `updowncounter`) |
| `k8s.statefulset.current_pods` (type: `gauge`) | `k8s.statefulset.current_pods` (type: `updowncounter`) |
| `k8s.statefulset.updated_pods` (type: `gauge`) | `k8s.statefulset.updated_pods` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s HorizontalPodAutoscaler metrics
The K8s HorizontalPodAutoscaler metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1636](https://github.com/open-telemetry/semantic-conventions/pull/1636) (TODO: replace with SemConv version once
available).
The changes in their metric names and types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|------------------------------------------------|
| `k8s.hpa.desired_replicas` (type: `gauge`) | `k8s.hpa.desired_pods` (type: `updowncounter`) |
| `k8s.hpa.current_replicas` (type: `gauge`) | `k8s.hpa.current_pods` (type: `updowncounter`) |
| `k8s.hpa.max_replicas` (type: `gauge`) | `k8s.hpa.max_pods` (type: `updowncounter`) |
| `k8s.hpa.min_replicas` (type: `gauge`) | `k8s.hpa.min_pods` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s DaemonSet metrics
The K8s DaemonSet metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1649](https://github.com/open-telemetry/semantic-conventions/pull/1649) (TODO: replace with SemConv version once
available).
The changes in their metric types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|-----------------------------------------------------------------|
| `k8s.daemonset.current_scheduled_nodes` (type: `gauge`) | `k8s.daemonset.current_scheduled_nodes` (type: `updowncounter`) |
| `k8s.daemonset.desired_scheduled_nodes` (type: `gauge`) | `k8s.daemonset.desired_scheduled_nodes` (type: `updowncounter`) |
| `k8s.daemonset.misscheduled_nodes` (type: `gauge`) | `k8s.daemonset.misscheduled_nodes` (type: `updowncounter`) |
| `k8s.daemonset.ready_nodes` (type: `gauge`) | `k8s.daemonset.ready_nodes` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s Job metrics
The K8s Job metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1649](https://github.com/open-telemetry/semantic-conventions/pull/1660) (TODO: replace with SemConv version once
available).
The changes in their metric types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|----------------------------------------------------------|----------------------------------------|
| `k8s.job.active_pods` (type: `gauge`) | `k8s.job.active_pods` (type: `updowncounter`) |
| `k8s.job.failed_pods` (type: `gauge`) | `k8s.job.failed_pods` (type: `updowncounter`) |
| `k8s.job.desired_successful_pods` (type: `gauge`) | `k8s.job.desired_successful_pods` (type: `updowncounter`) |
| `k8s.job.max_parallel_pods` (type: `gauge`) | `k8s.job.max_parallel_pods` (type: `updowncounter`) |
### K8s Cronjob metrics
The K8s Cronjob metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1649](https://github.com/open-telemetry/semantic-conventions/pull/1660) (TODO: replace with SemConv version once
available).
The changes in their metric types are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|--------------------------------------------------|--------------------------------|
| `k8s.cronjob.active_jobs` (type: `gauge`) | `k8s.cronjob.active_jobs` (type: `updowncounter`) |
<!-- prettier-ignore-end -->
### K8s Namespace metrics
The K8s Namespace metrics implemented by the Collector and specifically the
[k8scluster](https://github.com/open-telemetry/opentelemetry-collector-contrib/blob/v0.115.0/receiver/k8sclusterreceiver/documentation.md)
receiver were introduced as semantic conventions in
[#1649](https://github.com/open-telemetry/semantic-conventions/pull/1668) (TODO: replace with SemConv version once
available).
The changes in their metrics are the following:
<!-- prettier-ignore-start -->
| Old (Collector) ![changed](https://img.shields.io/badge/changed-orange?style=flat) | New |
|------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------|
| `k8s.namespace.phase` (type: `gauge`), 1 for active and 0 for terminating | `k8s.namespace.phase` (type: `updowncounter`), with the attribute `k8s.namespace.phase` indicating the phase |
<!-- prettier-ignore-end -->