mirror of
https://github.com/VictoriaMetrics/VictoriaMetrics.git
synced 2024-11-23 20:37:12 +01:00
9ded04e643
* docs: remove raw and endraw tags as they are not needed for the new version of site Signed-off-by: Artem Navoiev <tenmozes@gmail.com> * revert formating in vmaler Signed-off-by: Artem Navoiev <tenmozes@gmail.com> --------- Signed-off-by: Artem Navoiev <tenmozes@gmail.com>
89 lines
3.4 KiB
Markdown
89 lines
3.4 KiB
Markdown
---
|
|
sort: 31
|
|
weight: 31
|
|
title: VictoriaMetrics Cluster Per Tenant Statistic
|
|
menu:
|
|
docs:
|
|
parent: 'victoriametrics'
|
|
weight: 31
|
|
aliases:
|
|
- /PerTenantStatistic.html
|
|
---
|
|
|
|
# VictoriaMetrics Cluster Per Tenant Statistic
|
|
|
|
***The per-tenant statistic is a part of [enterprise package](https://docs.victoriametrics.com/enterprise.html). It is available for download and evaluation at [releases page](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/latest). You need to request a [free trial license](https://victoriametrics.com/products/enterprise/trial/) for evaluation.***
|
|
|
|
<img alt="cluster-per-tenant-stat" src="PerTenantStatistic-stats.webp">
|
|
|
|
VictoriaMetrics cluster for enterprise provides various metrics and statistics usage per tenant:
|
|
|
|
- `vminsert`
|
|
- `vm_tenant_inserted_rows_total` - total number of inserted rows. Find out which tenant
|
|
puts the most of the pressure on the storage.
|
|
|
|
- `vmselect`
|
|
- `vm_tenant_select_requests_duration_ms_total` - query latency.
|
|
Helps to identify tenants with the heaviest queries.
|
|
- `vm_tenant_select_requests_total` - total number of requests.
|
|
Discover which tenant sends the most of the queries and how it changes with time.
|
|
|
|
- `vmstorage`
|
|
- `vm_tenant_active_timeseries` - number of active time series.
|
|
This metric correlates with memory usage, so can be used to find the most expensive
|
|
tenant in terms of memory.
|
|
- `vm_tenant_used_tenant_bytes` - disk space usage. Helps to track disk space usage
|
|
per tenant.
|
|
- `vm_tenant_timeseries_created_total` - number of new time series created. Helps to track
|
|
the churn rate per tenant, or identify inefficient usage of the system.
|
|
|
|
Collect the metrics by any scrape agent you like (`vmagent`, `victoriametrics`, Prometheus, etc) and put into TSDB.
|
|
It is ok to use existing cluster for storing such metrics, but make sure to use a different tenant for it to avoid collisions.
|
|
Or just run a separate TSDB (VM single, Prometheus, etc.) to keep the data isolated from the main cluster.
|
|
|
|
Example of the scraping configuration for statistic is the following:
|
|
|
|
```yaml
|
|
scrape_configs:
|
|
- job_name: cluster
|
|
scrape_interval: 10s
|
|
static_configs:
|
|
- targets: ['vmselect:8481','vmstorage:8482','vminsert:8480']
|
|
```
|
|
|
|
## Visualization
|
|
|
|
Use [this dashboard](https://grafana.com/grafana/dashboards/16399) for visualizing per-tenant statistics in Grafana.
|
|
|
|
## Integration with vmgateway
|
|
|
|
`vmgateway` supports integration with Per Tenant Statistics data for rate limiting purposes.
|
|
More information can be found [here](https://docs.victoriametrics.com/vmgateway.html)
|
|
|
|
## Integration with vmalert
|
|
|
|
You can generate alerts based on each tenant's resource usage and send notifications
|
|
to prevent limits exhaustion.
|
|
|
|
Here is an alert example for high churn rate by the tenant:
|
|
|
|
```yaml
|
|
|
|
- alert: TooHighChurnRate
|
|
expr: |
|
|
(
|
|
sum(rate(vm_tenant_timeseries_created_total[5m])) by(accountID,projectID)
|
|
/
|
|
sum(rate(vm_tenant_inserted_rows_total[5m])) by(accountID,projectID)
|
|
) > 0.1
|
|
for: 15m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: "Churn rate is more than 10% for the last 15m"
|
|
description: "VM constantly creates new time series in the tenant: {{ $labels.accountID }}:{{ $labels.projectID }}.\n
|
|
This effect is known as Churn Rate.\n
|
|
High Churn Rate is tightly connected with database performance and may
|
|
result in unexpected OOM's or slow queries."
|
|
```
|