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
| `features`<br>`OTEL_EBPF_METRICS_FEATURES` | The list of metric groups OBI exports data for, refer to [metrics export features](#metrics-export-features). Accepted values `application`, `application_span`, `application_host`, `application_service_graph`, `network` and `network_inter_zone`. | list of strings | `["application"]` |
30
+
| YAML<br>environment variable | Description | Type | Default |
| `features`<br>`OTEL_EBPF_METRICS_FEATURES` | The list of metric groups OBI exports data for, refer to [metrics export features](#metrics-export-features). Accepted values `all`, `*`, `application`, `application_span`, `application_host`, `application_service_graph`, `network` and `network_inter_zone`. | list of strings | `["application"]` |
33
33
34
34
### Metrics export features
35
35
36
36
The OBI metrics exporter can export the following metrics data groups for
37
37
processes matching entries in the [metrics discovery](./) configuration.
38
38
39
+
- `all` or `*`: All metric groups (convenience option for enabling all metrics)
39
40
- `application`: Application-level metrics.
40
41
- `application_host`: Application-level host metrics for host-based pricing.
41
42
- `application_span`: Application-level trace span metrics in legacy format
@@ -157,6 +158,7 @@ The list of instrumentation areas OBI can collection data from:
157
158
- `mqtt`: MQTT publish/subscribe message metrics (MQTT 3.1.1 and 5.0)
158
159
- `couchbase`: Couchbase N1QL/SQL++ query metrics and KV (Key-Value) protocol
159
160
metrics based on memcached protocol
161
+
- `genai`: GenAI client metrics (OpenAI and Anthropic)
160
162
- `gpu`: GPU performance metrics
161
163
- `mongo`: MongoDB client call metrics
162
164
- `dns`: DNS query metrics
@@ -216,6 +218,7 @@ The list of instrumentation areas OBI can collection data from:
216
218
- `mqtt`: MQTT publish/subscribe message traces (MQTT 3.1.1 and 5.0)
217
219
- `couchbase`: Couchbase N1QL/SQL++ query traces and KV (Key-Value) protocol
218
220
traces, with query text and operation details
221
+
- `genai`: GenAI client traces (OpenAI and Anthropic)
219
222
- `gpu`: GPU performance traces
220
223
- `mongo`: MongoDB client call traces
221
224
- `dns`: DNS query traces
@@ -419,23 +422,24 @@ pull metrics in Prometheus format. It is enabled if the `port` property is set.
| `port`<br>`OTEL_EBPF_PROMETHEUS_PORT` | The HTTP port for the Prometheus scrape endpoint. If unset or 0, no Prometheus endpoint is open. | int | |
433
-
| `path`<br>`OTEL_EBPF_PROMETHEUS_PATH` | The HTTP query path to fetch the list of Prometheus metrics. | string | `/metrics` |
434
-
| `extra_resource_attributes`<br>`OTEL_EBPF_PROMETHEUS_EXTRA_RESOURCE_ATTRIBUTES` | A list of additional resource attributes to be added to the reported `target_info` metric. Refer to [extra resource attributes](#prometheus-extra-resource-attributes) for important details about runtime discovered attributes. | list of strings | |
435
-
| `ttl`<br>`OTEL_EBPF_PROMETHEUS_TTL` | The duration after which metric instances are not reported if they haven't been updated. Used to avoid reporting indefinitely finished application instances. | Duration | `5m` |
436
-
| `buckets` | Sets how you can override bucket boundaries of diverse histograms, refer to [override histogram buckets](../metrics-histograms/). | Object | |
437
-
| `allow_service_graph_self_references`<br>`OTEL_EBPF_PROMETHEUS_ALLOW_SERVICE_GRAPH_SELF_REFERENCES` | Does OBI include self-referencing service in service graph generation. Self referencing isn't useful for service graphs and increases data cardinality. | boolean | `false` |
438
-
| `instrumentations`<br>`OTEL_EBPF_PROMETHEUS_INSTRUMENTATIONS` | The list of instrumentation OBI collects data for, refer to [Prometheus instrumentation](#prometheus-instrumentation) section. | list of strings | `["*"]` |
433
+
| YAML<br>environment variable | Description | Type | Default |
| `port`<br>`OTEL_EBPF_PROMETHEUS_PORT` | The HTTP port for the Prometheus scrape endpoint. If unset or 0, no Prometheus endpoint is open. | int | |
436
+
| `path`<br>`OTEL_EBPF_PROMETHEUS_PATH` | The HTTP query path to fetch the list of Prometheus metrics. | string | `/metrics` |
437
+
| `extra_resource_attributes`<br>`OTEL_EBPF_PROMETHEUS_EXTRA_RESOURCE_ATTRIBUTES` | A list of additional resource attributes to be added to the reported `target_info` metric. Refer to [extra resource attributes](#prometheus-extra-resource-attributes) for important details about runtime discovered attributes. | list of strings | |
438
+
| `ttl`<br>`OTEL_EBPF_PROMETHEUS_TTL` | The duration after which metric instances are not reported if they haven't been updated. Used to avoid reporting indefinitely finished application instances. | Duration | `5m` |
439
+
| `buckets` | Sets how you can override bucket boundaries of diverse histograms, refer to [override histogram buckets](../metrics-histograms/). | Object | |
440
+
| `exemplar_filter`<br>`OTEL_EBPF_PROMETHEUS_EXEMPLAR_FILTER` | Controls when exemplars are attached to Prometheus metrics. Accepted values: `always_on`, `always_off`, `trace_based`. | string | `always_off` |
441
+
| `allow_service_graph_self_references`<br>`OTEL_EBPF_PROMETHEUS_ALLOW_SERVICE_GRAPH_SELF_REFERENCES` | Does OBI include self-referencing service in service graph generation. Self referencing isn't useful for service graphs and increases data cardinality. | boolean | `false` |
442
+
| `instrumentations`<br>`OTEL_EBPF_PROMETHEUS_INSTRUMENTATIONS` | The list of instrumentation OBI collects data for, refer to [Prometheus instrumentation](#prometheus-instrumentation) section. | list of strings | `["*"]` |
439
443
440
444
### Prometheus extra resource attributes
441
445
@@ -468,6 +472,7 @@ The list of instrumentation areas OBI can collection data from:
0 commit comments