Kinesis and Cortex Integration

Input and output integration overview

The Kinesis plugin enables you to read from Kinesis data streams, supporting various data formats and configurations.

This plugin enables Telegraf to send metrics to Cortex using the Prometheus remote write protocol, allowing seamless ingestion into Cortex’s scalable, multi-tenant time series storage.

Integration details

Kinesis

The Kinesis Telegraf plugin is designed to read from Amazon Kinesis data streams, enabling users to gather metrics in real-time. As a service input plugin, it operates by listening for incoming data rather than polling at regular intervals. The configuration specifies various options including the AWS region, stream name, authentication credentials, and data formats. It supports tracking of undelivered messages to prevent data loss, and users can utilize DynamoDB for maintaining checkpoints of the last processed records. This plugin is particularly useful for applications requiring reliable and scalable stream processing alongside other monitoring needs.

Cortex

With Telegraf’s HTTP output plugin and the prometheusremotewrite data format you can send metrics directly to Cortex, a horizontally scalable, long-term storage backend for Prometheus. Cortex supports multi-tenancy and accepts remote write requests using the Prometheus protobuf format. By using Telegraf as the collection agent and Remote Write as the transport mechanism, organizations can extend observability into sources not natively supported by Prometheus—such as Windows hosts, SNMP-enabled devices, or custom application metrics—while leveraging Cortex’s high-availability and long-retention capabilities.

Configuration

Kinesis

# Configuration for the AWS Kinesis input.
[[inputs.kinesis_consumer]]
  ## Amazon REGION of kinesis endpoint.
  region = "ap-southeast-2"

  ## Amazon Credentials
  ## Credentials are loaded in the following order
  ## 1) Web identity provider credentials via STS if role_arn and web_identity_token_file are specified
  ## 2) Assumed credentials via STS if role_arn is specified
  ## 3) explicit credentials from 'access_key' and 'secret_key'
  ## 4) shared profile from 'profile'
  ## 5) environment variables
  ## 6) shared credentials file
  ## 7) EC2 Instance Profile
  # access_key = ""
  # secret_key = ""
  # token = ""
  # role_arn = ""
  # web_identity_token_file = ""
  # role_session_name = ""
  # profile = ""
  # shared_credential_file = ""

  ## Endpoint to make request against, the correct endpoint is automatically
  ## determined and this option should only be set if you wish to override the
  ## default.
  ##   ex: endpoint_url = "http://localhost:8000"
  # endpoint_url = ""

  ## Kinesis StreamName must exist prior to starting telegraf.
  streamname = "StreamName"

  ## Shard iterator type (only 'TRIM_HORIZON' and 'LATEST' currently supported)
  # shard_iterator_type = "TRIM_HORIZON"

  ## Max undelivered messages
  ## This plugin uses tracking metrics, which ensure messages are read to
  ## outputs before acknowledging them to the original broker to ensure data
  ## is not lost. This option sets the maximum messages to read from the
  ## broker that have not been written by an output.
  ##
  ## This value needs to be picked with awareness of the agent's
  ## metric_batch_size value as well. Setting max undelivered messages too high
  ## can result in a constant stream of data batches to the output. While
  ## setting it too low may never flush the broker's messages.
  # max_undelivered_messages = 1000

  ## Data format to consume.
  ## Each data format has its own unique set of configuration options, read
  ## more about them here:
  ## https://github.com/influxdata/telegraf/blob/master/docs/DATA_FORMATS_INPUT.md
  data_format = "influx"

  ##
  ## The content encoding of the data from kinesis
  ## If you are processing a cloudwatch logs kinesis stream then set this to "gzip"
  ## as AWS compresses cloudwatch log data before it is sent to kinesis (aws
  ## also base64 encodes the zip byte data before pushing to the stream.  The base64 decoding
  ## is done automatically by the golang sdk, as data is read from kinesis)
  ##
  # content_encoding = "identity"

  ## Optional
  ## Configuration for a dynamodb checkpoint
  [inputs.kinesis_consumer.checkpoint_dynamodb]
    ## unique name for this consumer
    app_name = "default"
    table_name = "default"

Cortex

[[outputs.http]]
  ## Cortex Remote Write endpoint
  url = "http://cortex.example.com/api/v1/push"

  ## Use POST to send data
  method = "POST"

  ## Send metrics using Prometheus remote write format
  data_format = "prometheusremotewrite"

  ## Optional HTTP headers for authentication
  # [outputs.http.headers]
  #   X-Scope-OrgID = "your-tenant-id"
  #   Authorization = "Bearer YOUR_API_TOKEN"

  ## Optional TLS configuration
  # tls_ca = "/path/to/ca.pem"
  # tls_cert = "/path/to/cert.pem"
  # tls_key = "/path/to/key.pem"
  # insecure_skip_verify = false

  ## Request timeout
  timeout = "10s"

Input and output integration examples

Kinesis

1. Real-Time Data Processing with Kinesis: This use case involves integrating the Kinesis plugin with a monitoring dashboard to analyze incoming data metrics in real-time. For instance, an application could consume logs from multiple services and present them visually, allowing operations teams to quickly identify trends and react to anomalies as they occur.

2. Serverless Log Aggregation: Utilize this plugin in a serverless architecture where Kinesis streams aggregate logs from various microservices. The plugin can create metrics that help detect issues in the system, automating alerting processes through third-party integrations, enabling teams to minimize downtime and improve reliability.

3. Dynamic Scaling Based on Stream Metrics: Implement a solution where stream metrics consumed by the Kinesis plugin could be used to adjust resources dynamically. For example, if the number of records processed spikes, corresponding scale-up actions could be triggered to handle the increased load, ensuring optimal resource allocation and performance.

4. Data Pipeline to S3 with Checkpointing: Create a robust data pipeline where Kinesis stream data is processed through the Telegraf Kinesis plugin, with checkpoints stored in DynamoDB. This approach can ensure data consistency and reliability, as it manages the state of processed data, enabling seamless integration with downstream data lakes or storage solutions.

Cortex

1. Unified Multi-Tenant Monitoring: Use Telegraf to collect metrics from different teams or environments and push them to Cortex with separate X-Scope-OrgID headers. This enables isolated data ingestion and querying per tenant, ideal for managed services and platform teams.

2. Extending Prometheus Coverage to Edge Devices: Deploy Telegraf on edge or IoT devices to collect system metrics and send them to a centralized Cortex cluster. This approach ensures consistent observability even for environments without local Prometheus scrapers.

3. Global Service Observability with Federated Tenants: Aggregate metrics from global infrastructure by configuring Telegraf agents to push data into regional Cortex clusters, each tagged with tenant identifiers. Cortex handles deduplication and centralized access across regions.

4. Custom App Telemetry Pipeline: Collect app-specific telemetry via Telegraf’s exec or http input plugins and forward it to Cortex. This allows DevOps teams to monitor app-specific KPIs in a scalable, query-efficient format while keeping metrics logically grouped by tenant or service.

Feedback

Thank you for being part of our community! If you have any general feedback or found any bugs on these pages, we welcome and encourage your input. Please submit your feedback in the InfluxDB community Slack.