Lagrande is a load-testing tool for your TSDB. It is written in Go and it was designed to support many data formats, sending over UDP/TCP or HTTP and being highly configurable in regards to the metrics it's generating.
- Integer fixed value
- Float fixed value
- Integer counter (increment/decrement)
- Float counter (increment/decrement)
- Integer random value
- Float random value
- Latency (random float generated from a Beta probability distribution)
First start Graphite and Grafana containers so that you have a destination to send metrics to and a nice dashboard to plot your metrics:
docker run --rm -d --name graphite -p 80:80 -p 2003-2004:2003-2004 -p 8125:8125/udp -p 8126:8126 graphiteapp/graphite-statsd
docker run --rm -d --name grafana -p 3000:3000 grafana/grafanaThen grab one of the pre-built releases, move it to one of the folders in your $PATH and just launch it: lagrande.
Pre-built binairies will be made available as GitHub releases.
Make sure your local Golang installation is properly set up and then run:
go get github.com/aleveille/lagrande
go install github.com/aleveille/lagrandeThe command line arguments listed below are also documented with the usual Go -h flag. Only the most important flags will be documented here:
| Flag | Default value | Accepted values | Description |
|---|---|---|---|
-endpoint |
<empty> |
<URI string> |
The endpoint to send the data to, must be an URI. |
-format |
carbon |
carbon, influxdb, atlas |
The data format. Some TSDB support more than one format. |
-protocol |
auto |
auto, http, tcp, udp |
Auto will automatically pick an appropriate protocol based on the format (eg: HTTP for Atlas and TCP for Carbon) Not all formats support all protocols! |
-profile |
'counterInt={name: fixedValue, value: 10, increment: 0},randomInt={name: jiggle, min: 50, max: 75} |
<string> |
The configuration for the generator(s) to use. See Metric generation reference. |
-interval |
1s |
<Go duration string> |
How often each worker will generate metrics. |
-metricNamespacePrefix |
lagrande. |
<string> |
For namespacing metrics, this will be prepended to the metric name. Support placeholders: NODENAME, WORKERNUM, WORKERFULLNAME |
-metricNamespaceSuffix |
-WORKERNUM |
<string> |
For namespacing metrics, this will be appended to the metric name. Support placeholders: NODENAME, WORKERNUM, WORKERFULLNAME |
-tags |
node=NODENAME,process=lagrande,thread=WORKERFULLNAME |
<string> |
Comma-delimited list of tags of format name=value. Supports placeholders: NODENAME, PID, WORKERNUM, WORKERFULLNAME, METRICNAME. |
-workersCount |
10 |
<URI> |
Number of parallel workers that will send metrics. |
-workersInterval |
1s |
<Go duration string> |
Wait time between starting workers, must be a >= 0 Go Duration. |
You can use -profile flag to provide inline configuration on which generators to create and how to configure them.
The general syntax is generatorType={[key: value][, ...]}[, ...]. Spaces are optional, but help in reading the configuration.
Eg: counterInt={name: metric1, value: 10}, randomInt={name: random, min: 10, max: 20}
| Generator type | Config elements |
|---|---|
| counterInt |
|
| counterFloat |
|
| latency |
|
| randomInt |
|
| randomFloat |
|
This profile will create a metric with a value of 42 that never changes. It doesn't require any computation and is therefore very fast. A fixed value counterInt generator can be very good if you're interested in generating a very high number of metrics and you aren't much interested in their variance.
lagrande -profile 'counterInt={name: staticValue, value: 42, increment: 0}'
Integer counters with a low cardinality (<500000) gets their string byte representation cached. This counter is also very efficient. When this counter gets to 100000, it will reset its value to 0.
lagrande -profile 'counterInt={name: counter, value: 0, increment: 1, maximum: 100000}'
By using the reset: false configuration parameter for the generator, you can have the counter plateau at its final value.
lagrande -profile 'counterInt={name: biggerCounter, value: 0, increment: 1, reset: false}'
The increment can be a negative number. This counter will decrease from 1000.00 to 0.00 and will stay at 0.00 when it gets there.
lagrande -profile 'counterFloat={name: floatCountdown, value: 1000, increment: -2.5, minimum: 0, reset: false}'
Random int can be good to simulate discrete values such as the number of users on a website. This is a gauge-type metric.
lagrande -profile 'randomInt={name: connectedUsers, min: 10, max: 200}'
Random float can be good to simulate continuous values such as percentages. This is a gauge-type metric.
lagrande -profile 'randomFloat={name: someBufferUsage, min: 0, max: 1}'
The latency generator generates a random number between its min and max configuration parameters, but the generation is done with a gamma probability distribution. This means most values will be close to the lower bound (min parameter) and higher values have a decreased chance of being generated.
lagrande -profile 'latency={name: requestTime, min: 150, max: 8000, alpha: 1.5, beta: 10}'
It is possible more than one generator. The number of metrics per seconds will be: (Number of workers * Number of generator) / Interval.
lagrande -profile 'counterInt={name: staticValue, value: 42, increment: 0}, counterInt={name: counter, value: 0, increment: 1, maximum: 100000}, randomInt={name: connectedUsers, min: 10, max: 200}, randomFloat={name: someBufferUsage, min: 0, max: 1}'
TSDB:
- Graphite:
docker run --rm -d --name graphite -p 80:80 -p 2003-2004:2003-2004 -p 8125:8125/udp -p 8126:8126 graphiteapp/graphite-statsd - InfluxDB:
docker run --rm -d --name influxdb -p 8086:8086 influxdb - IRONdb:
docker run --rm -d --name irondb -p 2003-2004:2003-2004 -p 8112:8112 irondb/irondb - M3DB:
docker run --rm -d --name m3db -p 7201:7201 -p 7203:7203 -p 9003:9003 -v $HOME/tmp/m3db_data:/var/lib/m3db --privileged quay.io/m3db/m3dbnode:latest - Timescale:
docker run --rm -d --name timescaledb -p 5432:5432 -e POSTGRES_PASSWORD=password timescale/timescaledb:latest-pg11 - VictoriaMetrics:
docker run --rm -d --name victoriametrics -p 2003:2003 -p 8428:8428 victoriametrics/victoria-metrics -graphiteListenAddr :2003
Support tooling:
- Grafana:
docker run --rm -d --name grafana -p 3000:3000 grafana/grafana
If you have a feature request or a question, feel free to open an issue. Otherwise, I accept contributions via GitHub pull requests.
MIT License
Copyright (c) 2019 Alexandre Léveillé
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.