[Yandex Cloud documentation](../../index.md) > [Monium](../index.md) > [Data delivery](index.md) > OTLP in Monium # OTLP in Monium Monium can receive telemetry via [OTLP](https://opentelemetry.io/docs/specs/otlp/) (OpenTelemetry Protocol), which allows you to send data directly from applications using the OpenTelemetry SDK or through OpenTelemetry Collector. ## Endpoints {#endpoints} Monium provides the following endpoints to receive telemetry in OTLP format: #| || **Protocol** | **Endpoint** | **Description** || || HTTP | `https://ingest.monium.yandex.cloud/otlp/v1/metrics` | HTTP endpoint for receiving metrics || || HTTP | `https://ingest.monium.yandex.cloud/otlp/v1/logs` | HTTP endpoint for receiving logs || || HTTP | `https://ingest.monium.yandex.cloud/otlp/v1/traces` | HTTP endpoint for receiving traces || || gRPC | `ingest.monium.yandex.cloud:443` | gRPC endpoint for receiving all telemetry || |# The gRPC and HTTP endpoints are fully compatible with the [OTLP specification](https://opentelemetry.io/docs/specs/otlp/). ## Authorization {#authorization} Requests are authorized using an API key provided in the `Authorization` header: ``` Authorization: Api-Key ``` The API key is created for a service account with the `monium.telemetry.writer` role. The key must have the `yc.monium.telemetry.write` scope to write any type of telemetry or one or more of the following for granular access: `yc.monium.metrics.write`, `yc.monium.logs.write`, `yc.monium.traces.write`. [Learn more about creating an API key](../../iam/operations/iam-token/create-for-sa.md). ## Headings {#headers} ### Required headers {#required-headers} #| || **Header** | **Description** | **Example** || || `Authorization` | API key for authentication | `Api-Key AQVN...` || || `x-monium-project` | Project name: Monium | `folder__b1...` || |# ### Optional headers {#optional-headers} #| || **Header** | **Description** | **Example** | **Default value** || || `x-monium-cluster` | Cluster name | `production` | `default` || || `x-monium-service` | Service name | `payment-gateway` | `default` || |# ## Distribution by shards {#shard-distribution} All metrics, logs, and traces in Monium have mandatory labels: `project`, `cluster`, and `service`. These labels form the shard key. When delivering telemetry in OpenTelemetry format, the values of these key attributes are assigned in the following order of priority: * Values assigned via HTTP or gRPC headers have the highest priority. * Then follow the values set in the resource attributes of the request body with the `cluster` and `service` keys. * Next are the values set in the resource attributes recommended by the OpenTelemetry semantic convention. * If the above-listed values are not set, the default value is used. The project name is only taken from the header. The algorithm for determining the shard key is presented in the table below. #| || header | resource’s own attribute | resource’s standard attribute | default value || || `x-monium-project` | — | — | — || || `x-monium-cluster` | `cluster` | `deployment.name` | `default` || || `x-monium-service` | `service` | `service.name`, `k8s.deployment.name`, `k8s.namespace.name` | `default` || |# Learn more about sharding parameters in [Basic terms](../concepts/glossary.md#shard). ## Data compression {#compression} Realtime API supports the following compression types: #| || **Method** | **HTTP header** | **gRPC** | **Recommendation** || || No compression. | – | – | For debugging or for a small data stream || || gzip | `Content-Encoding: gzip` | `grpc-encoding: gzip` | Supported by default || || zstd | `Content-Encoding: zstd` | `grpc-encoding: zstd` | **Recommended**: more efficient in terms of computing resources || |# ## Metric type mapping {#metric-type-mapping} This section maps OpenTelemetry data types to Monium types. ### OTLP metric types {#otlp-types} OpenTelemetry uses the following metric types: #| || **OTLP type** | **Characteristic** | **Use cases** || || `Gauge` | Instant value | Memory usage, temperature || || `Counter` | Monotonic counter | Number of requests, request size || || `UpDownCounter` | Additive counter | Number of active connections || || `Histogram` | Distribution of values | Request duration, response size || || `Summary` (deprecated) | Quantiles and sum | Latency percentiles || |# ### Monium type mapping {#monium-mapping} Monium converts OTLP types into its internal metric types depending on temporality: #### Temporality = Delta {#delta-temporality} #| || **OTLP type** | **Monotonicity** | **Monium type** | **Note** || || `Gauge` | — | `GAUGE` | Instant value || || `Sum` | Monotonic | `RATE` | The sum is divided by the interval length in seconds || || `Histogram` | — | `HIST_RATE` for buckets, `GAUGE` for statistics | `name={metric}, bin={bound}` + `name={metric}.count` + `name={metric}.sum` + `name={metric}.min` + `name={metric}.max` || || `Summary` | — | `GAUGE` | `name={metric}.quantile, quantile={q}` + `name={metric}.count` + `name={metric}.sum` + `name={metric}.min` + `name={metric}.max` || |# #### Temporality = Cumulative {#cumulative-temporality} #| || **OTLP type** | **Monotonicity** | **Monium type** | **Note** || || `Gauge` | — | `GAUGE` | Instant value || || `Sum` | Non-monotonic | `GAUGE` | Absolute value of the counter || || `Sum` | Monotonic | `GAUGE` | Absolute value of the counter || || `Histogram` | — | `HIST` for buckets, `GAUGE` for statistics | `name={metric}, bin={bound}` + `name={metric}.count` + `name={metric}.sum` + `name={metric}.min` + `name={metric}.max` || || `Summary` | — | `GAUGE` | `name={metric}.quantile, quantile={q}` + `name={metric}.count` + `name={metric}.sum` + `name={metric}.min` + `name={metric}.max` || |# ## Examples of sum processing {#sum-examples} ### Cumulative Sum {#cumulative-sum-example} #| || **Period** | **Increments** | **Sum** | **Value in Monium** | **Type** | **Note** || || #1 | `[1,1,1,2,2,2,3,3]` | 15 | 15 | `GAUGE` | The sum value is used || || #2 | `[3,4,1,2,-15]` | 10 | 10 | `GAUGE` | The sum value is used || || #3 | `[]` | 10 | 10 | `GAUGE` | The last value is saved || |# ### Delta Sum, Monotonic {#delta-sum-monotonic-example} #| || **Period** | **Increments** | **Sum, interval** | **Value in Monium** | **Type** | **Note** || || #1 | `[1,1,1,2,2,2,3,3]` | `15, (0, 5]` | 3 (15 / 5) | `RATE` | Sum is divided by interval || || #2 | `[3,4,1,2]` | `10, (5, 10]` | 2 (10 / 5) | `RATE` | Sum is divided by interval || || #3 | `[]` | `0, (10, 15]` | 0 (0 / 5) | `RATE` | Null value || |# Non-monotonic delta sums are typically not used in OpenTelemetry. ## Processing resource attributes {#resource-attributes} Monium processes OTLP resource attributes as follows: ### Preserved attributes {#preserved-attributes} The following resource attributes are preserved as metric labels: **Host attributes** (prioritized from highest to lowest): * `host` * `host.name` * `hostname` These attributes are saved under the `host` label. **Environment attributes** (prioritized from highest to lowest): * `env` * `deployment.environment.name` * `deployment.environment` These attributes are saved under the `env` label. **Kubernetes and Cloud attributes** (saved without renaming): * `cloud.availability_zone` * `cloud.region` * `container.name` * `k8s.cluster.name` * `k8s.container.name` * `k8s.cronjob.name` * `k8s.daemonset.name` * `k8s.deployment.name` * `k8s.job.name` * `k8s.namespace.name` * `k8s.pod.name` * `k8s.replicaset.name` * `k8s.statefulset.name` ### Reserved labels {#reserved-labels} If data point attributes use reserved label names, they are automatically renamed: #| || **Original name** | **Renamed to** || || `project` | `_project_` || || `cluster` | `_cluster_` || || `service` | `_service_` || |# ### Other attributes {#other-attributes} Resource attributes not listed above are discarded. All data point attributes are preserved as Monium metric labels. ## Data format {#data-format} Monium supports the **OTLP Protobuf** format for all types of telemetry. The OTLP JSON format is currently supported for logs and traces. ## See also {#see-also} * [SDKs for data transmission in OpenTelemetry format](otlp-sdk.md) * [Data transmission via OpenTelemetry Collector](opentelemetry.md) * [OTLP specification](https://opentelemetry.io/docs/specs/otlp/) * [OpenTelemetry Metrics](https://opentelemetry.io/docs/concepts/signals/metrics/)