You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/reference/edot-collector/config/configure-profiles-collection.md
+35-16Lines changed: 35 additions & 16 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,6 +1,6 @@
1
1
---
2
2
navigation_title: Profiles collection
3
-
description: Configure the EDOT Collector for profiles collection.
3
+
description: Learn how to configure and customize profiles collection through the Elastic Distribution of OpenTelemetry Collector.
4
4
applies_to:
5
5
stack: preview 9.2
6
6
serverless:
@@ -12,26 +12,27 @@ products:
12
12
- id: edot-collector
13
13
---
14
14
15
-
# Profiles support in EDOT
15
+
# Configure profiles collection
16
16
17
-
Profiles represent the fourth pillar of observability, complementing logs, traces, and metrics.
17
+
The {{edot}} (EDOT) Collector includes a profiling receiver, which offers an eBPF-based, system-wide profiler.
18
18
19
-
The Elastic Distributions of OpenTelemetry (EDOT) Collector includes a profiling receiver, which offers an eBPF-based, system-wide profiler.
19
+
To activate and configure profiling and send profiles to {{ecloud}} or {{es}}, follow these instructions.
20
20
21
-
Deploying the EDOT Collector with a profiling receiver in Kubernetes should be done as a DaemonSet. This ensures comprehensive, node-level profiling across the entire complete cluster, providing consistent data collection, resilience, scalability, and simplified management. This approach is recommended for optimal performance and complete observability.
21
+
:::{important}
22
+
OpenTelemetry profiling is still under active development. Refer to [The State of Profiling](https://opentelemetry.io/blog/2024/state-profiling/) blog post for more information.
23
+
:::
22
24
23
-
::{note}
24
-
OpenTelemetry profiles is not yet a stable feature. To use it, Elastic Distributions
25
-
of OpenTelemetry (EDOT) Collector need to be started with the additional argument `--feature-gates=service.profilesSupport`.
26
-
::
25
+
## Turn on profiling
27
26
28
-
## Full profiles collection
27
+
Follow these steps to turn on profiles collection through the EDOT Collector.
29
28
30
-
For full profiles collection and reporting use the following configuration.
29
+
:::::{stepper}
30
+
::::{step} Set up Universal Profiling
31
+
Follow the steps to [`set up Universal Profiling`] before applying the following configuration.
32
+
::::
31
33
32
-
::{note}
33
-
Please follow the steps in Kibana to [`set up Universal Profiling`] before applying the following configuration.
34
-
::
34
+
::::{step} Full profiles collection
35
+
For full profiles collection and reporting, use this Collector configuration:
35
36
36
37
```yaml
37
38
receivers:
@@ -44,10 +45,24 @@ service:
44
45
receivers: [ profiling ]
45
46
exporters: [ elasticsearch ]
46
47
```
48
+
::::
49
+
50
+
::::{step} Activate profiling in the Collector
51
+
To activate profiling in the EDOT Collector, start it using the additional argument `--feature-gates=service.profilesSupport`.
The system can be configured to generate and report metrics exclusively from profile information. This method contributes to a reduction in ingest traffic and storage costs.
63
+
You can configure the components to generate and report metrics exclusively from profile information. This method contributes to a reduction in ingest traffic and storage costs.
64
+
65
+
The following example generates profiling metrics by frame, frame type, and classification:
51
66
52
67
```yaml
53
68
connectors:
@@ -70,4 +85,8 @@ service:
70
85
exporters: [ elasticsearch ]
71
86
```
72
87
73
-
[`set up Universal Profiling`]: https://www.elastic.co/docs/solutions/observability/infra-and-hosts/get-started-with-universal-profiling#profiling-configure-data-ingestion
88
+
## Kubernetes deployments
89
+
90
+
In Kubernetes, deploy the EDOT Collector with a profiling receiver as a DaemonSet. This ensures comprehensive, node-level profiling across the entire complete cluster, providing consistent data collection, resilience, scalability, and simplified management. This approach is recommended for optimal performance and full observability.
91
+
92
+
[`set up Universal Profiling`]: docs-content://solutions/observability/infra-and-hosts/get-started-with-universal-profiling.md#profiling-configure-data-ingestion
0 commit comments