Bananabacon

Why Bananabacon?

Did you know you can create fake bacon from banana peel? This project creates fake logs and metrics for testing observability solutions.

What does it do?

It serves metrics on /metrics and replays a given log file by printing it to stdout. The metrics are generated using JavaScript expressions that run in Goja. Expressions are evaluated every time the /metrics endpoint is called. The log output is generated from a reference file, where the first log line's timestamp is mapped to the current time and the following log lines are output with a delay so that the replay is in-sync with the original output. That means when there was a 10 second gap between two log lines in the original log, the lines will be output with a 10 seconds gap as well.

How do I use it?

Control Bananabacon using the following general environment variables:

Variable	Description	Default
INPUT_FILE	The log file to replay.	/logs/test.log
FILTER_REGEX	The regex for filtering log lines.	.*
TIME_REGEX	The regex for extracting the timestamp from a log line. Must have exactly one subgroup for the timestamp.	(None)
TIME_FORMAT	The format of the timestamp in the logs, defined in the Go time format.	(None)
LOOP	Whether to loop the log output after the file has been replayed.	false
METRICS_PORT	Port the metrics server listens on.	8080

Add metrics to produce using the following environment variables (<name> stands for the exported metric name):

Variable	Description	Default
METRIC_<name>_EXPR	The expression generating the metric value. For counter this needs to return an int, for gauge any number. The variable t holds the passed milliseconds since the server was started, prev holds the last emitted value (or null in the first call). You can either provide a function: function (t, prev) { return t * 2 } or an expression: t * 2	t. Check below for examples for different metric types.
METRIC_<name>_TYPE	The metric type (counter, gauge, histogram, summary, untyped).	counter
METRIC_<name>_DESCR	The description for the metric that will be printed in the HELP line	""
METRIC_<name>_LABEL	The labels for the metric in the format key1=value1,key2=value2,key3=value3.	(None)

Example: Counter metric named my_metric sloping up and then becoming static.

METRIC_my_metric_EXPR = function(t, prev) { if(t < 100000) return t / 1000; return 100; }
METRIC_my_metric_TYPE = counter
METRIC_my_metric_DESCR = Sloping up until 100, then staying there
METRIC_my_metric_LABEL = my_app=app

This will produce:

# HELP Sloping up until 100, then staying there
# TYPE counter
my_metric {my_app="app"} 100

Example: Counter metric named my_metric counting up whenever it is evaluated

METRIC_my_metric_EXPR = (prev || 0) + 1
METRIC_my_metric_TYPE = counter
METRIC_my_metric_DESCR = Counting up
METRIC_my_metric_LABEL = my_app=app

This will produce:

# HELP Counting up
# TYPE counter
my_metric {my_app="app"} 1

Example: Gauge metric named my_metric emitting random values.

METRIC_my_metric_EXPR = Math.random()
METRIC_my_metric_TYPE = gauge
METRIC_my_metric_DESCR = Random values
METRIC_my_metric_LABEL = my_app=app

This will produce:

# HELP Random values
# TYPE gauge
my_metric {my_app="app"} 0.3434543

Example: Histogram metric named my_metric emitting a static histogram.

METRIC_my_metric_EXPR = ({"1.0": 1, "2.0": 2, "3.0": 4, "sum": 5, "count": 6})
METRIC_my_metric_TYPE = histogram
METRIC_my_metric_DESCR = Static histogram
METRIC_my_metric_LABEL = my_app=app

This will produce:

# HELP Static histogram
# TYPE histogram
my_metric_bucket {my_app="app", le="1.0"} 1
my_metric_bucket {my_app="app", le="2.0"} 2
my_metric_bucket {my_app="app", le="3.0"} 4
my_metric_sum {my_app="app"} 5
my_metric_count {my_app="app"} 6

Example: Summary metric named my_metric emitting static quantiles.

METRIC_my_metric_EXPR = ({"1.0": 1, "2.0": 2, "3.0": 4})
METRIC_my_metric_TYPE = summary
METRIC_my_metric_DESCR = Static summary
METRIC_my_metric_LABEL = my_app=app

This will produce:

# HELP Static summary
# TYPE summary
my_metric {my_app="app", quantile="1.0"} 1
my_metric {my_app="app", quantile="2.0"} 2
my_metric {my_app="app", quantile="3.0"} 4

Running with Docker

docker build -t bananabacon .
docker run -v /absolute/path/to/log:/logs/test.log -e TIME_REGEX='(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\.\d{3}).*' -e METRIC_my_metric_EXPR='(prev || 0) + 1' -p 8080:8080 bananabacon

TODOs

• More tests