Enrich your observability data from a CSV

A guide to using the new CSV enrichment feature

We’re excited to share that we’ve released a new feature that enables users to enrich events flowing through the topology using a CSV file.

Enrichment tables are a new concept in Vector that enables you to enrich events from external data sources. To start, we’ve added the ability to enrich events from a CSV file by looking up a row, or rows, matching provided conditions, allowing users to map the data into the event using the full power of VRL.

To support mapping events based on enrichment table data, two new VRL functions are now available:

For example, when collecting events from IoT devices, you may want to keep your payloads coming from the devices to be small. By enriching events from a CSV file, users can reformat the data to be more human readable and provide better context (e.g., converting data emitted by the IoT device — 1, 2, 3 — to "Low battery", "Medium battery", "High battery").

Let’s stick with the IoT example from above, and let’s assume that our CSV file contains the below:

1,"device battery full"
2,"device battery good"
3,"device battery ok"
4,"device battery low"
5,"device battery critical"

In order to use the csv file (let’s call it iot_remap.csv), the following Vector configuration is required:

type = "file"

path = "/etc/vector/iot_remap.csv"
encoding = { type = "csv" }

code = "integer"
message = "string"

Now, to translate the output from IoT devices to human-readable messages in our iot_remap.csv we can make use of the get_enrichment_table_record function:

type = "remap"
inputs = ["vector_agents"]
source = '''
. = parse_json!(.message)

code = del(.code)

row = get_enrichment_table_record!("iot_remap", { "code":  code })
.message = row.message

For our next steps, we’ll look to add support for or conditions and add additional enrichment table types (e.g., reading from Redis), but if you any feedback in the meantime, let us know on our Discord chat or Twitter.