Vector

Vector is a lightweight and ultra-fast tool for building observability pipelines. It has a built-in support for shipping logs to Axiom through the axiom sink.

Installation

Follow the quickstart guide in the Vector documentation to install Vector, and to configure sources and sinks.

timestamp field

Logs and metrics sent via vector should use @timestamp as the timestamp field like: {"@timestamp":"2022-04-14T21:30:30.658Z..."}, not _time. Axiom accepts many date strings and timestamps without knowing the format in advance, including Unix Epoch, RFC3339, and ISO 8601.

Configuration

Send data to Axiom with Vector using the file method and the axiom sink.

The example below configures Vector to read and collect logs from files and send them to Axiom:

  1. Create a dataset in Axiom.

  2. Generate an Axiom API token.

  3. Create a vector configuration file vector.toml with the following content:

    [sources.VECTOR_SOURCE_ID]
    type = "file"
    include = ["PATH_TO_LOGS"]
    
    [sinks.SINK_ID]
    type = "axiom"
    inputs = ["VECTOR_SOURCE_ID"]
    token = "AXIOM_API_TOKEN"
    dataset = "AXIOM_DATASET"
    
  4. In the code above, replace the following:

    • Replace VECTOR_SOURCE_ID with the Vector source ID.
    • Replace PATH_TO_LOGS with the path to the log files. For example, /var/log/**/*.log.
    • Replace SINK_ID with the sink ID.
    • Replace AXIOM_API_TOKEN with the Axiom API token you have generated.
    • Replace AXIOM_DATASET with the name of the Axiom dataset where you want to send data.
  5. Run Vector to send logs to Axiom.

Example with data transformation

The example below deletes a field before sending the data to Axiom:

[sources.VECTOR_SOURCE_ID]
type = "file"
include = ["PATH_TO_LOGS"]

[transforms.filter_json_fields]
type = "remap"
inputs = ["VECTOR_SOURCE_ID"]
source = '''
  . = del(.FIELD_TO_REMOVE)

[sinks.SINK_ID]
type = "axiom"
inputs = ["filter_json_fields"]
token = "AXIOM_API_TOKEN"
dataset = "AXIOM_DATASET"

Replace FIELD_TO_REMOVE with the field you want to remove.

Any changes to Vector’s file method can make the code example above outdated. If this happens, please refer to the official Vector documentation on the file method, and we kindly ask you to inform us of the issue using the feedback tool at the bottom of this page.

Send Kubernetes logs to Axiom

Send Kubernetes logs to Axiom using the Kubernetes source.

[sources.my_source_id]
type = "kubernetes_logs"
auto_partial_merge = true
ignore_older_secs = 600
read_from = "beginning"
self_node_name = "${VECTOR_SELF_NODE_NAME}"
exclude_paths_glob_patterns = [ "**/exclude/**" ]
extra_field_selector = "metadata.name!=pod-name-to-exclude"
extra_label_selector = "my_custom_label!=my_value"
extra_namespace_label_selector = "my_custom_label!=my_value"
max_read_bytes = 2_048
max_line_bytes = 32_768
fingerprint_lines = 1
glob_minimum_cooldown_ms = 60_000
delay_deletion_ms = 60_000
data_dir = "/var/lib/vector"
timezone = "local"

[sinks.axiom]
type = "axiom"
inputs = ["my_source_id"]
token = "xaat-1234"
dataset = "vector-dev"

DATASET is the name of your dataset. When logs are sent from your vector, it’s stored in a dataset in Axiom.

See creating a dataset for more

TOKEN is used to ingest or query data to your dataset. API token can be generated from settings on Axiom dashboard.

See creating an API token for more

Send Docker logs to Axiom

To send Docker logs using the Axiom sink, you need to create a configuration file, for example, vector.toml, with the following content:

# Define the Docker logs source
[sources.docker_logs]
type = "docker_logs"
docker_host = "unix:///var/run/docker.sock"

# Define the Axiom sink
[sinks.axiom]
type = "axiom"
inputs = ["docker_logs"]
dataset = "your_dataset_name"  # replace with the name of your Axiom dataset
token = "your_api_token"  # replace with your Axiom API token

Replace your_dataset_name with the name of the dataset you want to send logs to in Axiom, and your_api_token with your Axiom API token.

  • Run Vector: Start Vector with the configuration file you just created:
vector --config /path/to/vector.toml

Vector collects logs from Docker and forward them to Axiom using the Axiom sink. You can view and analyze your logs in your dataset.

Send AWS S3 logs to Axiom

To send AWS S3 logs using the Axiom sink, create a configuration file, for example, vector.toml, with the following content:

[sources.my_s3_source]
type = "aws_s3"
bucket = "my-bucket"  # replace with your bucket name
region = "us-west-2"  # replace with the AWS region of your bucket

  
[sinks.axiom]
type = "axiom"
inputs = ["my_s3_source"]
dataset = "your_dataset_name"  # replace with the name of your Axiom dataset
token = "your_api_token"  # replace with your Axiom API token

Finally, run Vector with the configuration file using vector --config ./vector.toml. This starts Vector and begins reading logs from the specified S3 bucket and sending them to the specified Axiom dataset.

Send Kafka logs to Axiom

To send Kafka logs using the Axiom sink, you need to create a configuration file, for example, vector.toml, with the following code:

[sources.my_kafka_source]
type = "kafka" # must be: kafka
bootstrap_servers = "10.14.22.123:9092" # your Kafka bootstrap servers
group_id = "my_group_id" # your Kafka consumer group ID
topics = ["my_topic"] # the Kafka topics to consume from
auto_offset_reset = "earliest" # start reading from the beginning

[sinks.axiom]
type = "axiom"
inputs = ["my_kafka_source"]  # connect the Axiom sink to your Kafka source
dataset = "your_dataset_name"  # replace with the name of your Axiom dataset
token = "your_api_token"  # replace with your Axiom API token

Finally, you can start Vector with your configuration file: vector --config /path/to/your/vector.toml

Send NGINX metrics to Axiom

To send NGINX metrics using Vector to the Axiom sink, first enable NGINX to emit metrics, then use Vector to capture and forward those metrics. Here is a step-by-step guide:

Step 1: Enable NGINX Metrics

Configure NGINX to expose metrics. This typically involves enabling the ngx_http_stub_status_module module in your NGINX configuration.

  1. Open your NGINX configuration file (often located at /etc/nginx/nginx.conf) and in your server block, add:
location /metrics {
  stub_status;
  allow 127.0.0.1; # only allow requests from localhost
  deny all; # deny all other hosts
}
  1. Restart or reload NGINX to apply the changes:
sudo systemctl restart nginx

This exposes basic NGINX metrics at the /metrics endpoint on your server.

Step 2: Configure Vector

Configure Vector to scrape the NGINX metrics and send them to Axiom. Create a new configuration file (vector.toml), and add the following:

[sources.nginx_metrics]
type = "nginx_metrics" # must be: nginx_metrics
endpoints = ["http://localhost/metrics"] # the endpoint where NGINX metrics are exposed

[sinks.axiom]
type = "axiom" # must be: axiom
inputs = ["nginx_metrics"] # use the metrics from the NGINX source
dataset = "your_dataset_name"  # replace with the name of your Axiom dataset
token = "your_api_token"  # replace with your Axiom API token

Finally, you can start Vector with your configuration file: vector --config /path/to/your/vector.toml

Send Syslog logs to Axiom

To send Syslog logs using the Axiom sink, you need to create a configuration file, for example, vector.toml, with the following code:

[sources.my_source_id]
type="syslog"
address="0.0.0.0:6514"
max_length=102_400
mode="tcp"

[sinks.axiom]
type="axiom"
inputs = [ "my_source_id" ] # required
dataset="your_dataset_name" # replace with the name of your Axiom dataset 
token="your_api_token" # replace with your Axiom API token

Send Prometheus metrics to Axiom

To send Prometheus scrape metrics using the Axiom sink, you need to create a configuration file, for example, vector.toml, with the following code:

# Define the Prometheus source that scrapes metrics
[sources.my_prometheus_source]
type = "prometheus_scrape"  # scrape metrics from a Prometheus endpoint
endpoints = ["http://localhost:9090/metrics"]  # replace with your Prometheus endpoint

# Define Axiom sink where logs will be sent
[sinks.axiom]
type = "axiom"  # Axiom type
inputs = ["my_prometheus_source"]  # connect the Axiom sink to your Prometheus source
dataset = "your_prometheus_dataset"  # replace with the name of your Axiom dataset
token = "your_api_token"  # replace with your Axiom API token

Check out the advanced configuration on Batch, Buffer configuration, and Encoding on Vector Documentation