Prometheus remote write benchmarker (acts as a Prometheus sending node_exporter samples).
It can also run in scrape mode where it serves as a scrape server for Prometheus to ingest metrics from.
It uses the InfluxData node_exporter simulator to generate and progress time using realistic host metrics and cardinality.
The main controls for controlling cardinality is the number of hosts to generate metrics for. You can also use the new series percent to tweak how many host IDs churn every scrape cycle (which is why scrape interval is required, even when running in scrape mode).
You can also specify fixed labels to apply to each benchmarker instance should you want to query the metrics. Use the environment variable PROMREMOTEBENCH_LABELS_JSON
or the command line argument labels
.
Here is what a Kubernetes manifest looks like for hosting the benchmarker, simply turn up number of replicas for more load or targets.
apiVersion: apps/v1
kind: Deployment
metadata:
name: promremotebench
spec:
replicas: 2
selector:
matchLabels:
k8s-app: promremotebench
template:
metadata:
labels:
k8s-app: promremotebench
spec:
containers:
- name: promremotebench
image: quay.io/m3db/promremotebench:latest
env:
- name: PROMREMOTEBENCH_TARGET
# NB: this can be a comma separated string of targets for writing to multiple targets.
value: "http://m3coordinator-dedicated-bench-cluster:7201/api/v1/prom/remote/write"
- name: PROMREMOTEBENCH_NUM_HOSTS
value: "1000"
- name: PROMREMOTEBENCH_INTERVAL
value: "10"
- name: PROMREMOTEBENCH_BATCH
value: "128"
- name: PROMREMOTEBENCH_NEW_SERIES_PERCENTAGE
value: "0.01"
apiVersion: apps/v1
kind: Deployment
metadata:
name: promremotebench
spec:
replicas: 2
selector:
matchLabels:
k8s-app: promremotebench
template:
metadata:
labels:
k8s-app: promremotebench
spec:
containers:
- name: promremotebench
image: quay.io/m3db/promremotebench:latest
env:
- name: PROMREMOTEBENCH_SCRAPE_SERVER
value: "0.0.0.0:4242"
- name: PROMREMOTEBENCH_NUM_HOSTS
value: "1000"
- name: PROMREMOTEBENCH_NEW_SERIES_PERCENTAGE
value: "0.01"
ports:
- name: http
containerPort: 4242