New Reduce transform

Canonical Log Lines in Vector

Fan of Stripe’s Canonical Log Lines? We are too. You can now find a new Reduce! This allows you to turn a stream of many small events into a stream of less small events!

The fantastic article by Brandur Leach describes them best:

They’re a simple idea: in addition to their normal log traces, requests (or some other unit of work that’s executing) also emit one long log line at the end that pulls all its key telemetry into one place.

Here’s a canonical example:

[2019-03-18 22:48:32.999] canonical-log-line alloc_count=9123 auth_type=api_key database_queries=34 duration=0.009 http_method=POST http_path=/v1/charges http_status=200 key_id=mk_123 permissions_used=account_write rate_allowed=true rate_quota=100 rate_remaining=99 request_id=req_123 team=acquiring user_id=usr_123

Let’s build similar in Vector!

We’ll take a series of events:

{"timestamp": "...", "message": "Received GET /path", "request_id": "abcd1234", "request_path": "/path", "request_params": "..."}
{"timestamp": "...", "message": "Executed query in 5.2ms", "request_id": "abcd1234", "query": "SELECT * FROM table", "query_duration_ms": 5.2}
{"timestamp": "...", "message": "Rendered partial _partial.erb in 2.3ms", "request_id": "abcd1234", "template": "_partial.erb", "render_duration_ms": 2.3}
{"timestamp": "...", "message": "Executed query in 7.8ms", "request_id": "abcd1234", "query": "SELECT * FROM table", "query_duration_ms": 7.8}
{"timestamp": "...", "message": "Sent 200 in 15.2ms", "request_id": "abcd1234", "response_status": 200, "response_duration_ms": 5.2}

Then output this (but not formatted so nicely!):

  "timestamp_start": "...",
  "timestamp_end": "...",
  "request_id": "abcd1234",
  "request_path": "/path",
  "request_params": "...",
  "query_duration_ms": 13.0,
  "render_duration_ms": 2.3,
  "status": 200,
  "response_duration_ms": 5.2

We’ll run this config:

data_dir = "tmp"

  include = ["input.log"]
  start_at_beginning = true
  type = "file"
  fingerprinting.strategy = "device_and_inode"

  inputs = ["source0"]
  type = "json_parser"
  field = "message"

  inputs = ["transform0"]
  type = "reduce"
  identifier_fields = ["request_id"]
  ends_when.type = "check_fields"
  ends_when."response_status.exists" = true
  merge_strategies.message = "discard"
  merge_strategies.query = "discard"
  merge_strategies.template = "discard"
  merge_strategies.query_duration_ms = "sum"
  merge_strategies.render_duration_ms = "sum"
  merge_strategies.response_duration_ms = "sum"

  healthcheck = true
  inputs = ["transform1"]
  type = "file"
  path = "output.log"
  encoding = "ndjson"
  buffer.type = "memory"
  buffer.max_events = 500
  buffer.when_full = "block"s

We hope you find this useful!